| { | |
| "best_metric": 0.49329306040311344, | |
| "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-2/checkpoint-3207", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 3207, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 5.944660653119031e-06, | |
| "loss": 0.5558, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.157496901381872e-06, | |
| "loss": 0.5053, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.4907635748386383, | |
| "eval_matthews_correlation": 0.42854888313294803, | |
| "eval_runtime": 0.8054, | |
| "eval_samples_per_second": 1295.012, | |
| "eval_steps_per_second": 81.947, | |
| "step": 1069 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 4.370333149644712e-06, | |
| "loss": 0.4286, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 3.5831693979075514e-06, | |
| "loss": 0.3992, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.5015696883201599, | |
| "eval_matthews_correlation": 0.48409375114357234, | |
| "eval_runtime": 0.8234, | |
| "eval_samples_per_second": 1266.73, | |
| "eval_steps_per_second": 80.157, | |
| "step": 2138 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 2.796005646170392e-06, | |
| "loss": 0.3434, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 2.008841894433232e-06, | |
| "loss": 0.3371, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.5880187749862671, | |
| "eval_matthews_correlation": 0.49329306040311344, | |
| "eval_runtime": 0.9037, | |
| "eval_samples_per_second": 1154.15, | |
| "eval_steps_per_second": 73.033, | |
| "step": 3207 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4276, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 113339571995772.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 6.7318244048561916e-06, | |
| "num_train_epochs": 4, | |
| "per_device_train_batch_size": 8, | |
| "seed": 14 | |
| } | |
| } | |