| { | |
| "best_metric": 0.9615338526063512, | |
| "best_model_checkpoint": "bert-base-uncased-finetuned-spam-real/run-2/checkpoint-1670", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 2505, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 9.117546081542969, | |
| "learning_rate": 1.8393158320232495e-06, | |
| "loss": 0.4054, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9505692031156381, | |
| "eval_loss": 0.14854364097118378, | |
| "eval_runtime": 597.8505, | |
| "eval_samples_per_second": 27.917, | |
| "eval_steps_per_second": 1.746, | |
| "step": 835 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 1.4320781230926514, | |
| "learning_rate": 1.3806335796483742e-06, | |
| "loss": 0.1311, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 0.354846715927124, | |
| "learning_rate": 9.219513272734991e-07, | |
| "loss": 0.1227, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9615338526063512, | |
| "eval_loss": 0.14428438246250153, | |
| "eval_runtime": 598.1778, | |
| "eval_samples_per_second": 27.901, | |
| "eval_steps_per_second": 1.745, | |
| "step": 1670 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 2.810948371887207, | |
| "learning_rate": 4.6326907489862395e-07, | |
| "loss": 0.1123, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "grad_norm": 0.04734252020716667, | |
| "learning_rate": 4.586822523748751e-09, | |
| "loss": 0.0837, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9608148591971241, | |
| "eval_loss": 0.15702731907367706, | |
| "eval_runtime": 598.8615, | |
| "eval_samples_per_second": 27.87, | |
| "eval_steps_per_second": 1.743, | |
| "step": 2505 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2505, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 4990682275848000.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 2.2979980843981246e-06, | |
| "num_train_epochs": 3, | |
| "per_device_train_batch_size": 8, | |
| "seed": 24 | |
| } | |
| } | |