| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9335978527249388, | |
| "eval_steps": 250, | |
| "global_step": 1000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2333994631812347, | |
| "grad_norm": 3.116767168045044, | |
| "learning_rate": 8.969068649097766e-05, | |
| "loss": 0.2345, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.2333994631812347, | |
| "eval_loss": 0.12925492227077484, | |
| "eval_runtime": 44.2128, | |
| "eval_samples_per_second": 1.968, | |
| "eval_steps_per_second": 1.968, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.4667989263624694, | |
| "grad_norm": 2.796464681625366, | |
| "learning_rate": 5.79860922155139e-05, | |
| "loss": 0.0968, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4667989263624694, | |
| "eval_loss": 0.10360654443502426, | |
| "eval_runtime": 44.2406, | |
| "eval_samples_per_second": 1.967, | |
| "eval_steps_per_second": 1.967, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.7001983895437041, | |
| "grad_norm": 2.5315845012664795, | |
| "learning_rate": 2.2048719230965166e-05, | |
| "loss": 0.0845, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7001983895437041, | |
| "eval_loss": 0.0936887115240097, | |
| "eval_runtime": 43.1722, | |
| "eval_samples_per_second": 2.015, | |
| "eval_steps_per_second": 2.015, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.9335978527249388, | |
| "grad_norm": 2.071190118789673, | |
| "learning_rate": 1.2154116213611399e-06, | |
| "loss": 0.0766, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.9335978527249388, | |
| "eval_loss": 0.09068119525909424, | |
| "eval_runtime": 44.5694, | |
| "eval_samples_per_second": 1.952, | |
| "eval_steps_per_second": 1.952, | |
| "step": 1000 | |
| } | |
| ], | |
| "logging_steps": 250, | |
| "max_steps": 1071, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 250, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 9.76219910449582e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |