| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.6, | |
| "eval_steps": 5, | |
| "global_step": 20, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 1.3029956031514607, | |
| "learning_rate": 0.0001, | |
| "loss": 2.1603, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 1.3943801456437246, | |
| "learning_rate": 0.00018660254037844388, | |
| "loss": 1.8141, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_loss": 1.236501693725586, | |
| "eval_runtime": 3.059, | |
| "eval_samples_per_second": 6.538, | |
| "eval_steps_per_second": 1.635, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.8292993293983272, | |
| "learning_rate": 0.00011736481776669306, | |
| "loss": 1.163, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_loss": 1.0143945217132568, | |
| "eval_runtime": 1.8666, | |
| "eval_samples_per_second": 10.715, | |
| "eval_steps_per_second": 2.679, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.7648680289420796, | |
| "learning_rate": 3.5721239031346066e-05, | |
| "loss": 0.9305, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_loss": 0.9576635360717773, | |
| "eval_runtime": 1.8573, | |
| "eval_samples_per_second": 10.769, | |
| "eval_steps_per_second": 2.692, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 0.7309126849137305, | |
| "learning_rate": 0.0, | |
| "loss": 0.8644, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_loss": 0.9460625648498535, | |
| "eval_runtime": 1.868, | |
| "eval_samples_per_second": 10.707, | |
| "eval_steps_per_second": 2.677, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "step": 20, | |
| "total_flos": 842884448256.0, | |
| "train_loss": 1.2102977752685546, | |
| "train_runtime": 81.8583, | |
| "train_samples_per_second": 1.955, | |
| "train_steps_per_second": 0.244 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 20, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 842884448256.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |