| { | |
| "best_metric": 0.8788848263254113, | |
| "best_model_checkpoint": "berttiny-hate_speech18-bothpretrained\\run-40\\checkpoint-1916", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 1916, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 5.4840038745847375e-05, | |
| "loss": 0.4809, | |
| "step": 479 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.863345521023766, | |
| "eval_loss": 0.47785741090774536, | |
| "eval_runtime": 10.8777, | |
| "eval_samples_per_second": 201.145, | |
| "eval_steps_per_second": 1.655, | |
| "step": 479 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 3.6560025830564916e-05, | |
| "loss": 0.4595, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8752285191956124, | |
| "eval_loss": 0.4690297842025757, | |
| "eval_runtime": 10.868, | |
| "eval_samples_per_second": 201.324, | |
| "eval_steps_per_second": 1.656, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.8280012915282458e-05, | |
| "loss": 0.4477, | |
| "step": 1437 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8779707495429616, | |
| "eval_loss": 0.468118280172348, | |
| "eval_runtime": 10.9155, | |
| "eval_samples_per_second": 200.449, | |
| "eval_steps_per_second": 1.649, | |
| "step": 1437 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.4414, | |
| "step": 1916 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8788848263254113, | |
| "eval_loss": 0.46808430552482605, | |
| "eval_runtime": 10.8281, | |
| "eval_samples_per_second": 202.066, | |
| "eval_steps_per_second": 1.662, | |
| "step": 1916 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1916, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 4935921882816.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 7.312005166112983e-05, | |
| "num_train_epochs": 4, | |
| "per_device_train_batch_size": 16, | |
| "seed": 30 | |
| } | |
| } | |