| { | |
| "best_metric": 0.9048165137614679, | |
| "best_model_checkpoint": "bert-base-uncased-finetuned-sst2/run-3/checkpoint-5052", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 5052, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 7.637521743774414, | |
| "learning_rate": 3.0568943955117345e-06, | |
| "loss": 0.5094, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 104.65070343017578, | |
| "learning_rate": 2.721119527814927e-06, | |
| "loss": 0.4332, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 0.10155901312828064, | |
| "learning_rate": 2.3853446601181197e-06, | |
| "loss": 0.4318, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8899082568807339, | |
| "eval_loss": 0.4135288596153259, | |
| "eval_runtime": 2.3891, | |
| "eval_samples_per_second": 364.998, | |
| "eval_steps_per_second": 23.022, | |
| "step": 1684 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "grad_norm": 0.4222470819950104, | |
| "learning_rate": 2.0495697924213124e-06, | |
| "loss": 0.4108, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "grad_norm": 0.46430423855781555, | |
| "learning_rate": 1.7137949247245048e-06, | |
| "loss": 0.3485, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "grad_norm": 0.1579820215702057, | |
| "learning_rate": 1.3780200570276975e-06, | |
| "loss": 0.3125, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8967889908256881, | |
| "eval_loss": 0.4881651997566223, | |
| "eval_runtime": 2.45, | |
| "eval_samples_per_second": 355.915, | |
| "eval_steps_per_second": 22.449, | |
| "step": 3368 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 0.662672221660614, | |
| "learning_rate": 1.0422451893308901e-06, | |
| "loss": 0.3239, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "grad_norm": 10.589339256286621, | |
| "learning_rate": 7.064703216340827e-07, | |
| "loss": 0.2529, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "grad_norm": 0.06143077090382576, | |
| "learning_rate": 3.7069545393727534e-07, | |
| "loss": 0.2671, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.97, | |
| "grad_norm": 0.06450512260198593, | |
| "learning_rate": 3.492058624046797e-08, | |
| "loss": 0.2994, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9048165137614679, | |
| "eval_loss": 0.4830039441585541, | |
| "eval_runtime": 2.4197, | |
| "eval_samples_per_second": 360.376, | |
| "eval_steps_per_second": 22.73, | |
| "step": 5052 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5052, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 246789433818420.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 3.3926692632085418e-06, | |
| "num_train_epochs": 3, | |
| "per_device_train_batch_size": 4, | |
| "seed": 6 | |
| } | |
| } | |