| { | |
| "best_metric": 0.8325688073394495, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-25/checkpoint-527", | |
| "epoch": 8.0, | |
| "eval_steps": 500, | |
| "global_step": 4216, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 3.7773447036743164, | |
| "learning_rate": 0.0004316595412152798, | |
| "loss": 0.9448, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8325688073394495, | |
| "eval_loss": 1.1720139980316162, | |
| "eval_runtime": 2.8085, | |
| "eval_samples_per_second": 310.481, | |
| "eval_steps_per_second": 2.492, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 9.26517105102539, | |
| "learning_rate": 0.00036999389247023985, | |
| "loss": 0.4379, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8061926605504587, | |
| "eval_loss": 1.3765146732330322, | |
| "eval_runtime": 2.8111, | |
| "eval_samples_per_second": 310.197, | |
| "eval_steps_per_second": 2.49, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00030844525634141437, | |
| "loss": 0.3017, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8096330275229358, | |
| "eval_loss": 1.3439667224884033, | |
| "eval_runtime": 2.8182, | |
| "eval_samples_per_second": 309.415, | |
| "eval_steps_per_second": 2.484, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 4.918728828430176, | |
| "learning_rate": 0.0002467796075963744, | |
| "loss": 0.2219, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8073394495412844, | |
| "eval_loss": 1.4046818017959595, | |
| "eval_runtime": 2.8176, | |
| "eval_samples_per_second": 309.483, | |
| "eval_steps_per_second": 2.484, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 4.011919975280762, | |
| "learning_rate": 0.00018511395885133442, | |
| "loss": 0.1764, | |
| "step": 2635 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8153669724770642, | |
| "eval_loss": 1.4098650217056274, | |
| "eval_runtime": 2.8221, | |
| "eval_samples_per_second": 308.991, | |
| "eval_steps_per_second": 2.48, | |
| "step": 2635 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 8.373222351074219, | |
| "learning_rate": 0.00012344831010629444, | |
| "loss": 0.146, | |
| "step": 3162 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8153669724770642, | |
| "eval_loss": 1.367167592048645, | |
| "eval_runtime": 2.8214, | |
| "eval_samples_per_second": 309.068, | |
| "eval_steps_per_second": 2.481, | |
| "step": 3162 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 4.53749418258667, | |
| "learning_rate": 6.178266136125447e-05, | |
| "loss": 0.1245, | |
| "step": 3689 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.8153669724770642, | |
| "eval_loss": 1.4212154150009155, | |
| "eval_runtime": 2.8209, | |
| "eval_samples_per_second": 309.117, | |
| "eval_steps_per_second": 2.481, | |
| "step": 3689 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.7747430205345154, | |
| "learning_rate": 1.170126162144971e-07, | |
| "loss": 0.11, | |
| "step": 4216 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.819954128440367, | |
| "eval_loss": 1.392822027206421, | |
| "eval_runtime": 2.8243, | |
| "eval_samples_per_second": 308.751, | |
| "eval_steps_per_second": 2.479, | |
| "step": 4216 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4216, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 64687167744780.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.595362243351062, | |
| "learning_rate": 0.0004933251899603198, | |
| "num_train_epochs": 8, | |
| "temperature": 26 | |
| } | |
| } | |