| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 50.0, | |
| "global_step": 5750, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 7.35e-06, | |
| "loss": 11.7893, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 1.485e-05, | |
| "loss": 4.6989, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "learning_rate": 2.2349999999999998e-05, | |
| "loss": 3.5677, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 2.985e-05, | |
| "loss": 3.1589, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "eval_loss": 3.0830252170562744, | |
| "eval_runtime": 76.0167, | |
| "eval_samples_per_second": 22.363, | |
| "eval_steps_per_second": 1.408, | |
| "eval_wer": 1.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "learning_rate": 3.735e-05, | |
| "loss": 3.0279, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "learning_rate": 4.484999999999999e-05, | |
| "loss": 2.9807, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 6.09, | |
| "learning_rate": 5.234999999999999e-05, | |
| "loss": 2.9589, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 6.96, | |
| "learning_rate": 5.985e-05, | |
| "loss": 2.8921, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 6.96, | |
| "eval_loss": 2.6604607105255127, | |
| "eval_runtime": 76.2841, | |
| "eval_samples_per_second": 22.285, | |
| "eval_steps_per_second": 1.403, | |
| "eval_wer": 0.9981590808100045, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 7.83, | |
| "learning_rate": 6.735e-05, | |
| "loss": 2.346, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 8.7, | |
| "learning_rate": 7.484999999999999e-05, | |
| "loss": 1.6326, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 9.57, | |
| "learning_rate": 7.345263157894736e-05, | |
| "loss": 1.4075, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 10.43, | |
| "learning_rate": 7.187368421052631e-05, | |
| "loss": 1.3049, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 10.43, | |
| "eval_loss": 0.5069074630737305, | |
| "eval_runtime": 75.7661, | |
| "eval_samples_per_second": 22.437, | |
| "eval_steps_per_second": 1.412, | |
| "eval_wer": 0.5706849488986225, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 11.3, | |
| "learning_rate": 7.029473684210525e-05, | |
| "loss": 1.2407, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 12.17, | |
| "learning_rate": 6.87157894736842e-05, | |
| "loss": 1.1999, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "learning_rate": 6.713684210526315e-05, | |
| "loss": 1.167, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 13.91, | |
| "learning_rate": 6.55578947368421e-05, | |
| "loss": 1.1349, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 13.91, | |
| "eval_loss": 0.4158603847026825, | |
| "eval_runtime": 75.9244, | |
| "eval_samples_per_second": 22.391, | |
| "eval_steps_per_second": 1.409, | |
| "eval_wer": 0.5040944581984383, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 14.78, | |
| "learning_rate": 6.397894736842105e-05, | |
| "loss": 1.1161, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 15.65, | |
| "learning_rate": 6.239999999999999e-05, | |
| "loss": 1.0982, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 16.52, | |
| "learning_rate": 6.0821052631578945e-05, | |
| "loss": 1.0841, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 17.39, | |
| "learning_rate": 5.924210526315789e-05, | |
| "loss": 1.0686, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 17.39, | |
| "eval_loss": 0.38148975372314453, | |
| "eval_runtime": 77.84, | |
| "eval_samples_per_second": 21.84, | |
| "eval_steps_per_second": 1.375, | |
| "eval_wer": 0.4746397511585095, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 18.26, | |
| "learning_rate": 5.7663157894736836e-05, | |
| "loss": 1.047, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 19.13, | |
| "learning_rate": 5.608421052631579e-05, | |
| "loss": 1.03, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "learning_rate": 5.450526315789473e-05, | |
| "loss": 1.0144, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 20.87, | |
| "learning_rate": 5.292631578947368e-05, | |
| "loss": 0.999, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 20.87, | |
| "eval_loss": 0.3541375696659088, | |
| "eval_runtime": 75.1099, | |
| "eval_samples_per_second": 22.633, | |
| "eval_steps_per_second": 1.425, | |
| "eval_wer": 0.4343299688948137, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 21.74, | |
| "learning_rate": 5.1347368421052626e-05, | |
| "loss": 0.993, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 22.61, | |
| "learning_rate": 4.976842105263157e-05, | |
| "loss": 0.9766, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 23.48, | |
| "learning_rate": 4.8189473684210524e-05, | |
| "loss": 0.9573, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 24.35, | |
| "learning_rate": 4.661052631578947e-05, | |
| "loss": 0.945, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 24.35, | |
| "eval_loss": 0.3265579342842102, | |
| "eval_runtime": 75.4543, | |
| "eval_samples_per_second": 22.53, | |
| "eval_steps_per_second": 1.418, | |
| "eval_wer": 0.4131911381958992, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 25.22, | |
| "learning_rate": 4.5031578947368415e-05, | |
| "loss": 0.9309, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 26.09, | |
| "learning_rate": 4.345263157894736e-05, | |
| "loss": 0.9151, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 26.96, | |
| "learning_rate": 4.1873684210526314e-05, | |
| "loss": 0.9138, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 27.83, | |
| "learning_rate": 4.031052631578947e-05, | |
| "loss": 0.9058, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 27.83, | |
| "eval_loss": 0.2969067096710205, | |
| "eval_runtime": 75.0182, | |
| "eval_samples_per_second": 22.661, | |
| "eval_steps_per_second": 1.426, | |
| "eval_wer": 0.377071034088745, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 28.7, | |
| "learning_rate": 3.8731578947368425e-05, | |
| "loss": 0.9001, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 29.57, | |
| "learning_rate": 3.7152631578947364e-05, | |
| "loss": 0.8882, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 30.43, | |
| "learning_rate": 3.557368421052631e-05, | |
| "loss": 0.867, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 31.3, | |
| "learning_rate": 3.399473684210526e-05, | |
| "loss": 0.8672, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 31.3, | |
| "eval_loss": 0.2802095413208008, | |
| "eval_runtime": 75.6097, | |
| "eval_samples_per_second": 22.484, | |
| "eval_steps_per_second": 1.415, | |
| "eval_wer": 0.35529740366914236, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 32.17, | |
| "learning_rate": 3.241578947368421e-05, | |
| "loss": 0.8522, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 33.04, | |
| "learning_rate": 3.0836842105263153e-05, | |
| "loss": 0.8391, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 33.91, | |
| "learning_rate": 2.9257894736842103e-05, | |
| "loss": 0.8368, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 34.78, | |
| "learning_rate": 2.767894736842105e-05, | |
| "loss": 0.8313, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 34.78, | |
| "eval_loss": 0.26624372601509094, | |
| "eval_runtime": 74.7639, | |
| "eval_samples_per_second": 22.738, | |
| "eval_steps_per_second": 1.431, | |
| "eval_wer": 0.3379673712943566, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 35.65, | |
| "learning_rate": 2.6099999999999997e-05, | |
| "loss": 0.8188, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 36.52, | |
| "learning_rate": 2.4521052631578947e-05, | |
| "loss": 0.8168, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 37.39, | |
| "learning_rate": 2.2942105263157892e-05, | |
| "loss": 0.8103, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 38.26, | |
| "learning_rate": 2.136315789473684e-05, | |
| "loss": 0.8068, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 38.26, | |
| "eval_loss": 0.25275924801826477, | |
| "eval_runtime": 75.264, | |
| "eval_samples_per_second": 22.587, | |
| "eval_steps_per_second": 1.422, | |
| "eval_wer": 0.3180981400368184, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 39.13, | |
| "learning_rate": 1.9784210526315787e-05, | |
| "loss": 0.8046, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "learning_rate": 1.8205263157894736e-05, | |
| "loss": 0.7929, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 40.87, | |
| "learning_rate": 1.6626315789473682e-05, | |
| "loss": 0.7817, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 41.74, | |
| "learning_rate": 1.504736842105263e-05, | |
| "loss": 0.7796, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 41.74, | |
| "eval_loss": 0.25369659066200256, | |
| "eval_runtime": 74.2196, | |
| "eval_samples_per_second": 22.905, | |
| "eval_steps_per_second": 1.442, | |
| "eval_wer": 0.3073065447851203, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 42.61, | |
| "learning_rate": 1.3468421052631578e-05, | |
| "loss": 0.7789, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 43.48, | |
| "learning_rate": 1.1889473684210526e-05, | |
| "loss": 0.7736, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 44.35, | |
| "learning_rate": 1.0310526315789472e-05, | |
| "loss": 0.783, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 45.22, | |
| "learning_rate": 8.73157894736842e-06, | |
| "loss": 0.7621, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 45.22, | |
| "eval_loss": 0.2502739131450653, | |
| "eval_runtime": 74.1963, | |
| "eval_samples_per_second": 22.912, | |
| "eval_steps_per_second": 1.442, | |
| "eval_wer": 0.3035612264330604, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 46.09, | |
| "learning_rate": 7.168421052631579e-06, | |
| "loss": 0.767, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 46.96, | |
| "learning_rate": 5.589473684210526e-06, | |
| "loss": 0.7659, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 47.83, | |
| "learning_rate": 4.0105263157894735e-06, | |
| "loss": 0.7622, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 48.7, | |
| "learning_rate": 2.431578947368421e-06, | |
| "loss": 0.7611, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 48.7, | |
| "eval_loss": 0.24768942594528198, | |
| "eval_runtime": 74.3172, | |
| "eval_samples_per_second": 22.875, | |
| "eval_steps_per_second": 1.44, | |
| "eval_wer": 0.29905414841617467, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 49.57, | |
| "learning_rate": 8.526315789473684e-07, | |
| "loss": 0.7494, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "step": 5750, | |
| "total_flos": 2.9609555446330925e+19, | |
| "train_loss": 1.4450044761326002, | |
| "train_runtime": 12353.9156, | |
| "train_samples_per_second": 14.878, | |
| "train_steps_per_second": 0.465 | |
| } | |
| ], | |
| "max_steps": 5750, | |
| "num_train_epochs": 50, | |
| "total_flos": 2.9609555446330925e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |