| { | |
| "best_metric": 0.8371559633027523, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-38/checkpoint-527", | |
| "epoch": 6.0, | |
| "eval_steps": 500, | |
| "global_step": 3162, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 6.650190353393555, | |
| "learning_rate": 0.0003977381702368286, | |
| "loss": 0.8907, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8371559633027523, | |
| "eval_loss": 1.0675474405288696, | |
| "eval_runtime": 2.812, | |
| "eval_samples_per_second": 310.097, | |
| "eval_steps_per_second": 2.489, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 7.2698974609375, | |
| "learning_rate": 0.0003181905361894629, | |
| "loss": 0.4134, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8084862385321101, | |
| "eval_loss": 1.2476831674575806, | |
| "eval_runtime": 2.8048, | |
| "eval_samples_per_second": 310.9, | |
| "eval_steps_per_second": 2.496, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 14.755233764648438, | |
| "learning_rate": 0.00023864290214209717, | |
| "loss": 0.2802, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8119266055045872, | |
| "eval_loss": 1.312410831451416, | |
| "eval_runtime": 2.8236, | |
| "eval_samples_per_second": 308.82, | |
| "eval_steps_per_second": 2.479, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 4.266346454620361, | |
| "learning_rate": 0.00015909526809473144, | |
| "loss": 0.208, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8130733944954128, | |
| "eval_loss": 1.2842535972595215, | |
| "eval_runtime": 2.8057, | |
| "eval_samples_per_second": 310.791, | |
| "eval_steps_per_second": 2.495, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 4.0564045906066895, | |
| "learning_rate": 7.969857832449545e-05, | |
| "loss": 0.1663, | |
| "step": 2635 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.823394495412844, | |
| "eval_loss": 1.2945654392242432, | |
| "eval_runtime": 2.8126, | |
| "eval_samples_per_second": 310.037, | |
| "eval_steps_per_second": 2.489, | |
| "step": 2635 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 8.322790145874023, | |
| "learning_rate": 1.5094427712972622e-07, | |
| "loss": 0.1392, | |
| "step": 3162 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8176605504587156, | |
| "eval_loss": 1.3337725400924683, | |
| "eval_runtime": 2.8133, | |
| "eval_samples_per_second": 309.956, | |
| "eval_steps_per_second": 2.488, | |
| "step": 3162 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3162, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 6, | |
| "save_steps": 500, | |
| "total_flos": 48527917525620.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.63744345806538, | |
| "learning_rate": 0.00047728580428419435, | |
| "num_train_epochs": 6, | |
| "temperature": 29 | |
| } | |
| } | |