| { | |
| "best_metric": 0.634, | |
| "best_model_checkpoint": "./checkpoints/run-0/checkpoint-2500", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 3750, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 0.00018779999999999998, | |
| "loss": 0.6899, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 0.0002916, | |
| "loss": 0.6815, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 0.0002707333333333333, | |
| "loss": 0.6751, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.626, | |
| "eval_loss": 0.6668456792831421, | |
| "eval_runtime": 7.6525, | |
| "eval_samples_per_second": 653.38, | |
| "eval_steps_per_second": 81.672, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 0.00024986666666666665, | |
| "loss": 0.6697, | |
| "step": 1252 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 0.00022899999999999998, | |
| "loss": 0.6678, | |
| "step": 1565 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 0.00020813333333333334, | |
| "loss": 0.6671, | |
| "step": 1878 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "learning_rate": 0.00018726666666666665, | |
| "loss": 0.6634, | |
| "step": 2191 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.634, | |
| "eval_loss": 0.6575816869735718, | |
| "eval_runtime": 7.6029, | |
| "eval_samples_per_second": 657.642, | |
| "eval_steps_per_second": 82.205, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 0.00016639999999999998, | |
| "loss": 0.6587, | |
| "step": 2504 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 0.0001455333333333333, | |
| "loss": 0.6592, | |
| "step": 2817 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 0.00012466666666666667, | |
| "loss": 0.6576, | |
| "step": 3130 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 0.00010379999999999999, | |
| "loss": 0.6584, | |
| "step": 3443 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.6286, | |
| "eval_loss": 0.6556090712547302, | |
| "eval_runtime": 7.6842, | |
| "eval_samples_per_second": 650.684, | |
| "eval_steps_per_second": 81.335, | |
| "step": 3750 | |
| } | |
| ], | |
| "logging_steps": 313, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 69988590059520.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 0.0003, | |
| "per_device_train_batch_size": 16 | |
| } | |
| } | |