| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.006793478260869565, | |
| "eval_steps": 5, | |
| "global_step": 50, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0006793478260869565, | |
| "eval_loss": 4.499350547790527, | |
| "eval_runtime": 20.1523, | |
| "eval_samples_per_second": 975.424, | |
| "eval_steps_per_second": 7.642, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.001358695652173913, | |
| "eval_loss": 4.498117923736572, | |
| "eval_runtime": 20.4966, | |
| "eval_samples_per_second": 959.037, | |
| "eval_steps_per_second": 7.513, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.0020380434782608695, | |
| "eval_loss": 4.495996475219727, | |
| "eval_runtime": 21.0485, | |
| "eval_samples_per_second": 933.891, | |
| "eval_steps_per_second": 7.316, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.002717391304347826, | |
| "eval_loss": 4.492975234985352, | |
| "eval_runtime": 21.3506, | |
| "eval_samples_per_second": 920.677, | |
| "eval_steps_per_second": 7.213, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0033967391304347825, | |
| "eval_loss": 4.489028453826904, | |
| "eval_runtime": 21.5145, | |
| "eval_samples_per_second": 913.661, | |
| "eval_steps_per_second": 7.158, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.004076086956521739, | |
| "eval_loss": 4.484206199645996, | |
| "eval_runtime": 21.5836, | |
| "eval_samples_per_second": 910.738, | |
| "eval_steps_per_second": 7.135, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.004755434782608696, | |
| "eval_loss": 4.478429317474365, | |
| "eval_runtime": 21.7892, | |
| "eval_samples_per_second": 902.143, | |
| "eval_steps_per_second": 7.068, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.005434782608695652, | |
| "eval_loss": 4.471565246582031, | |
| "eval_runtime": 21.9727, | |
| "eval_samples_per_second": 894.608, | |
| "eval_steps_per_second": 7.009, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.006114130434782609, | |
| "eval_loss": 4.463589668273926, | |
| "eval_runtime": 21.6519, | |
| "eval_samples_per_second": 907.863, | |
| "eval_steps_per_second": 7.113, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.006793478260869565, | |
| "eval_loss": 4.454347610473633, | |
| "eval_runtime": 21.7784, | |
| "eval_samples_per_second": 902.593, | |
| "eval_steps_per_second": 7.071, | |
| "step": 50 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 22080, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 5, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |