| { | |
| "best_metric": 0.2534329295158386, | |
| "best_model_checkpoint": "./xls-r-300m-bemba-fullset/checkpoint-1000", | |
| "epoch": 1.9993275050437123, | |
| "global_step": 1486, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 1.47e-05, | |
| "loss": 11.7638, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 2.97e-05, | |
| "loss": 3.7497, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.4699999999999996e-05, | |
| "loss": 3.0287, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 5.97e-05, | |
| "loss": 2.8222, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 7.47e-05, | |
| "loss": 2.268, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "eval_loss": 1.0355395078659058, | |
| "eval_runtime": 140.0953, | |
| "eval_samples_per_second": 11.057, | |
| "eval_steps_per_second": 1.385, | |
| "eval_wer": 0.9928986442866365, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 8.969999999999998e-05, | |
| "loss": 1.1597, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 0.00010469999999999998, | |
| "loss": 0.9172, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "learning_rate": 0.0001197, | |
| "loss": 0.8284, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 0.0001347, | |
| "loss": 0.7667, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 0.00014969999999999998, | |
| "loss": 0.7365, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "eval_loss": 0.2534329295158386, | |
| "eval_runtime": 132.8836, | |
| "eval_samples_per_second": 11.657, | |
| "eval_steps_per_second": 1.46, | |
| "eval_wer": 0.8469980632666236, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 0.0001647, | |
| "loss": 0.7227, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 0.00017969999999999998, | |
| "loss": 0.7068, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "learning_rate": 0.0001947, | |
| "loss": 0.6875, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "learning_rate": 0.00020969999999999997, | |
| "loss": 0.6798, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1486, | |
| "total_flos": 4.4337322315078743e+18, | |
| "train_loss": 2.1141159319460474, | |
| "train_runtime": 3966.7132, | |
| "train_samples_per_second": 5.996, | |
| "train_steps_per_second": 0.375 | |
| } | |
| ], | |
| "max_steps": 1486, | |
| "num_train_epochs": 2, | |
| "total_flos": 4.4337322315078743e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |