{ "best_metric": 0.3379155099391937, "best_model_checkpoint": "../models/thchs_bl1k/checkpoint-252", "epoch": 4.064516129032258, "eval_steps": 126, "global_step": 252, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.016129032258064516, "grad_norm": Infinity, "learning_rate": 0.0, "loss": 4.8763, "step": 1 }, { "epoch": 1.0161290322580645, "grad_norm": 2.696150779724121, "learning_rate": 0.00062, "loss": 2.4002, "step": 63 }, { "epoch": 2.032258064516129, "grad_norm": 1.7860599756240845, "learning_rate": 0.0009528301886792453, "loss": 0.7877, "step": 126 }, { "epoch": 2.032258064516129, "eval_loss": 0.5588188767433167, "eval_runtime": 186.5879, "eval_samples_per_second": 1.067, "eval_steps_per_second": 0.134, "step": 126 }, { "epoch": 3.0483870967741935, "grad_norm": 1.0973650217056274, "learning_rate": 0.0008339622641509435, "loss": 0.4384, "step": 189 }, { "epoch": 4.064516129032258, "grad_norm": 0.6822451949119568, "learning_rate": 0.0007150943396226415, "loss": 0.3438, "step": 252 }, { "epoch": 4.064516129032258, "eval_loss": 0.3379155099391937, "eval_runtime": 74.5925, "eval_samples_per_second": 2.668, "eval_steps_per_second": 0.335, "step": 252 } ], "logging_steps": 63, "max_steps": 630, "num_input_tokens_seen": 0, "num_train_epochs": 11, "save_steps": 126, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 2.652896821248e+17, "train_batch_size": 16, "trial_name": null, "trial_params": null }