| { | |
| "best_metric": 18.448637316561843, | |
| "best_model_checkpoint": "./checkpoint-200", | |
| "epoch": 41.666666666666664, | |
| "global_step": 500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 5.746206252093564e-06, | |
| "loss": 1.0017, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "learning_rate": 7.226145833886759e-06, | |
| "loss": 0.1431, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "learning_rate": 8.045346441409077e-06, | |
| "loss": 0.0125, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "learning_rate": 8.614712701594175e-06, | |
| "loss": 0.0029, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "eval_loss": 0.6650390625, | |
| "eval_runtime": 18.7793, | |
| "eval_samples_per_second": 2.982, | |
| "eval_steps_per_second": 0.107, | |
| "eval_wer": 19.28721174004193, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 10.42, | |
| "learning_rate": 9.051535070126044e-06, | |
| "loss": 0.0014, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 12.5, | |
| "learning_rate": 9.406017565450707e-06, | |
| "loss": 0.0009, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 14.58, | |
| "learning_rate": 9.704332905698628e-06, | |
| "loss": 0.0007, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "learning_rate": 9.961869578736312e-06, | |
| "loss": 0.0005, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "eval_loss": 0.74560546875, | |
| "eval_runtime": 19.1155, | |
| "eval_samples_per_second": 2.93, | |
| "eval_steps_per_second": 0.105, | |
| "eval_wer": 18.448637316561843, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 18.75, | |
| "learning_rate": 9.333333333333334e-06, | |
| "loss": 0.0004, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 20.83, | |
| "learning_rate": 8.5e-06, | |
| "loss": 0.0004, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 22.92, | |
| "learning_rate": 7.666666666666667e-06, | |
| "loss": 0.0003, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "learning_rate": 6.833333333333334e-06, | |
| "loss": 0.0003, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_loss": 0.77978515625, | |
| "eval_runtime": 19.0324, | |
| "eval_samples_per_second": 2.942, | |
| "eval_steps_per_second": 0.105, | |
| "eval_wer": 19.49685534591195, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 27.08, | |
| "learning_rate": 6e-06, | |
| "loss": 0.0003, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 29.17, | |
| "learning_rate": 5.1666666666666675e-06, | |
| "loss": 0.0003, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 31.25, | |
| "learning_rate": 4.333333333333334e-06, | |
| "loss": 0.0002, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "learning_rate": 3.5e-06, | |
| "loss": 0.0002, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "eval_loss": 0.79638671875, | |
| "eval_runtime": 18.9232, | |
| "eval_samples_per_second": 2.959, | |
| "eval_steps_per_second": 0.106, | |
| "eval_wer": 19.70649895178197, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 35.42, | |
| "learning_rate": 2.666666666666667e-06, | |
| "loss": 0.0002, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 37.5, | |
| "learning_rate": 1.8333333333333333e-06, | |
| "loss": 0.0002, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 39.58, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 0.0002, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 41.67, | |
| "learning_rate": 1.6666666666666668e-07, | |
| "loss": 0.0002, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 41.67, | |
| "eval_loss": 0.80224609375, | |
| "eval_runtime": 19.0419, | |
| "eval_samples_per_second": 2.941, | |
| "eval_steps_per_second": 0.105, | |
| "eval_wer": 20.020964360587, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 41.67, | |
| "step": 500, | |
| "total_flos": 1.3379374861974818e+20, | |
| "train_loss": 0.05835451430082321, | |
| "train_runtime": 9110.0805, | |
| "train_samples_per_second": 7.025, | |
| "train_steps_per_second": 0.055 | |
| } | |
| ], | |
| "max_steps": 500, | |
| "num_train_epochs": 42, | |
| "total_flos": 1.3379374861974818e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |