{ "best_metric": null, "best_model_checkpoint": null, "epoch": 1.01, "eval_steps": 50, "global_step": 100, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.25, "grad_norm": 44.361637115478516, "learning_rate": 4.2000000000000006e-07, "loss": 2.441, "step": 25 }, { "epoch": 0.5, "grad_norm": 22.284318923950195, "learning_rate": 9.200000000000001e-07, "loss": 1.7682, "step": 50 }, { "epoch": 0.5, "eval_loss": 1.843565583229065, "eval_runtime": 1085.1769, "eval_samples_per_second": 2.994, "eval_steps_per_second": 0.375, "eval_wer": 32.099283342073065, "step": 50 }, { "epoch": 0.75, "grad_norm": 9.719245910644531, "learning_rate": 1.42e-06, "loss": 0.97, "step": 75 }, { "epoch": 1.01, "grad_norm": 7.1439056396484375, "learning_rate": 1.9200000000000003e-06, "loss": 0.5702, "step": 100 }, { "epoch": 1.01, "eval_loss": 1.1091302633285522, "eval_runtime": 1511.5514, "eval_samples_per_second": 2.149, "eval_steps_per_second": 0.269, "eval_wer": 55.43576297850026, "step": 100 }, { "epoch": 1.01, "step": 100, "total_flos": 9.2029884162048e+17, "train_loss": 1.437346134185791, "train_runtime": 3304.8553, "train_samples_per_second": 0.968, "train_steps_per_second": 0.03 } ], "logging_steps": 25, "max_steps": 100, "num_input_tokens_seen": 0, "num_train_epochs": 9223372036854775807, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 9.2029884162048e+17, "train_batch_size": 32, "trial_name": null, "trial_params": null }