| { | |
| "best_metric": 0.8268101761252447, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-3/checkpoint-288", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 288, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1.2690824270248413, | |
| "learning_rate": 0.00018636956997148543, | |
| "loss": 0.4818, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.786692759295499, | |
| "eval_f1": 0.8039568345323741, | |
| "eval_loss": 0.4252791106700897, | |
| "eval_precision": 0.7437603993344426, | |
| "eval_recall": 0.8747553816046967, | |
| "eval_runtime": 31.9977, | |
| "eval_samples_per_second": 31.94, | |
| "eval_steps_per_second": 1.0, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.9796446561813354, | |
| "learning_rate": 0.00015530797497623786, | |
| "loss": 0.4169, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7935420743639922, | |
| "eval_f1": 0.8117752007136486, | |
| "eval_loss": 0.4000055491924286, | |
| "eval_precision": 0.7459016393442623, | |
| "eval_recall": 0.8904109589041096, | |
| "eval_runtime": 30.9158, | |
| "eval_samples_per_second": 33.058, | |
| "eval_steps_per_second": 1.035, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 2.9656918048858643, | |
| "learning_rate": 0.00012424637998099028, | |
| "loss": 0.3978, | |
| "step": 288 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8268101761252447, | |
| "eval_f1": 0.8398190045248868, | |
| "eval_loss": 0.38850346207618713, | |
| "eval_precision": 0.7811447811447811, | |
| "eval_recall": 0.9080234833659491, | |
| "eval_runtime": 30.6883, | |
| "eval_samples_per_second": 33.303, | |
| "eval_steps_per_second": 1.043, | |
| "step": 288 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 672, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 7, | |
| "save_steps": 500, | |
| "total_flos": 707085591840.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.6895915406036343, | |
| "learning_rate": 0.000217431164966733, | |
| "num_train_epochs": 7, | |
| "temperature": 20 | |
| } | |
| } | |