| { | |
| "best_metric": 30.481593707691317, | |
| "best_model_checkpoint": "./checkpoint-200", | |
| "epoch": 38.012, | |
| "global_step": 500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 2.3000000000000004e-06, | |
| "loss": 1.7169, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 4.800000000000001e-06, | |
| "loss": 0.8922, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 5.02, | |
| "learning_rate": 7.3e-06, | |
| "loss": 0.4752, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 7.02, | |
| "learning_rate": 9.800000000000001e-06, | |
| "loss": 0.2693, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 7.02, | |
| "eval_loss": 0.6740689277648926, | |
| "eval_runtime": 1376.3341, | |
| "eval_samples_per_second": 0.732, | |
| "eval_steps_per_second": 0.023, | |
| "eval_wer": 35.48449487460512, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 9.02, | |
| "learning_rate": 9.425e-06, | |
| "loss": 0.1224, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 11.01, | |
| "learning_rate": 8.8e-06, | |
| "loss": 0.0482, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 13.01, | |
| "learning_rate": 8.175e-06, | |
| "loss": 0.0171, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 15.01, | |
| "learning_rate": 7.5500000000000006e-06, | |
| "loss": 0.0084, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 15.01, | |
| "eval_loss": 0.7756175398826599, | |
| "eval_runtime": 955.7345, | |
| "eval_samples_per_second": 1.055, | |
| "eval_steps_per_second": 0.033, | |
| "eval_wer": 30.481593707691317, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 17.01, | |
| "learning_rate": 6.925000000000001e-06, | |
| "loss": 0.0052, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 19.01, | |
| "learning_rate": 6.300000000000001e-06, | |
| "loss": 0.0039, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "learning_rate": 5.675000000000001e-06, | |
| "loss": 0.0033, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "learning_rate": 5.050000000000001e-06, | |
| "loss": 0.0029, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_loss": 0.8154371380805969, | |
| "eval_runtime": 1039.7317, | |
| "eval_samples_per_second": 0.969, | |
| "eval_steps_per_second": 0.031, | |
| "eval_wer": 31.4744374959706, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 24.03, | |
| "learning_rate": 4.425e-06, | |
| "loss": 0.0024, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 26.02, | |
| "learning_rate": 3.8000000000000005e-06, | |
| "loss": 0.0023, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 28.02, | |
| "learning_rate": 3.175e-06, | |
| "loss": 0.0021, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 30.02, | |
| "learning_rate": 2.55e-06, | |
| "loss": 0.002, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 30.02, | |
| "eval_loss": 0.8319515585899353, | |
| "eval_runtime": 954.6491, | |
| "eval_samples_per_second": 1.056, | |
| "eval_steps_per_second": 0.034, | |
| "eval_wer": 31.377731932177166, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 32.02, | |
| "learning_rate": 1.925e-06, | |
| "loss": 0.0019, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 34.02, | |
| "learning_rate": 1.3e-06, | |
| "loss": 0.0018, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 36.01, | |
| "learning_rate": 6.750000000000001e-07, | |
| "loss": 0.0018, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 38.01, | |
| "learning_rate": 5.0000000000000004e-08, | |
| "loss": 0.0017, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 38.01, | |
| "eval_loss": 0.8372375965118408, | |
| "eval_runtime": 984.3922, | |
| "eval_samples_per_second": 1.024, | |
| "eval_steps_per_second": 0.033, | |
| "eval_wer": 31.516343240281092, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 38.01, | |
| "step": 500, | |
| "total_flos": 8.59869058793472e+18, | |
| "train_loss": 0.17904520658403636, | |
| "train_runtime": 15686.2181, | |
| "train_samples_per_second": 2.04, | |
| "train_steps_per_second": 0.032 | |
| } | |
| ], | |
| "max_steps": 500, | |
| "num_train_epochs": 9223372036854775807, | |
| "total_flos": 8.59869058793472e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |