| { | |
| "best_metric": 0.8302752293577982, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-8/checkpoint-527", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 2635, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 10.257322311401367, | |
| "learning_rate": 0.0005386391253190077, | |
| "loss": 1.2539, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8302752293577982, | |
| "eval_loss": 1.5333763360977173, | |
| "eval_runtime": 2.8188, | |
| "eval_samples_per_second": 309.352, | |
| "eval_steps_per_second": 2.483, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.0004042348653959536, | |
| "loss": 0.5485, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.801605504587156, | |
| "eval_loss": 1.7583402395248413, | |
| "eval_runtime": 2.8148, | |
| "eval_samples_per_second": 309.788, | |
| "eval_steps_per_second": 2.487, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 25.133913040161133, | |
| "learning_rate": 0.0002695750840662017, | |
| "loss": 0.3506, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8096330275229358, | |
| "eval_loss": 1.8196579217910767, | |
| "eval_runtime": 2.8113, | |
| "eval_samples_per_second": 310.174, | |
| "eval_steps_per_second": 2.49, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 8.44449234008789, | |
| "learning_rate": 0.00013491530273644975, | |
| "loss": 0.2467, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8107798165137615, | |
| "eval_loss": 1.8072972297668457, | |
| "eval_runtime": 2.8135, | |
| "eval_samples_per_second": 309.934, | |
| "eval_steps_per_second": 2.488, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 8.500100135803223, | |
| "learning_rate": 2.555214066978215e-07, | |
| "loss": 0.19, | |
| "step": 2635 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8153669724770642, | |
| "eval_loss": 1.8087736368179321, | |
| "eval_runtime": 2.8195, | |
| "eval_samples_per_second": 309.277, | |
| "eval_steps_per_second": 2.483, | |
| "step": 2635 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2635, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 40447180739880.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.31906439494819705, | |
| "learning_rate": 0.0006732989066487597, | |
| "num_train_epochs": 5, | |
| "temperature": 15 | |
| } | |
| } | |