| { | |
| "best_metric": 0.879798903107861, | |
| "best_model_checkpoint": "berttiny-hate_speech18-bothpretrained\\run-37\\checkpoint-2874", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 3832, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 5.448730733370957e-05, | |
| "loss": 0.4754, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8642595978062158, | |
| "eval_loss": 0.4731012284755707, | |
| "eval_runtime": 10.8286, | |
| "eval_samples_per_second": 202.058, | |
| "eval_steps_per_second": 1.662, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 3.632487155580638e-05, | |
| "loss": 0.4543, | |
| "step": 1916 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8756855575868373, | |
| "eval_loss": 0.4699413478374481, | |
| "eval_runtime": 10.91, | |
| "eval_samples_per_second": 200.551, | |
| "eval_steps_per_second": 1.65, | |
| "step": 1916 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.816243577790319e-05, | |
| "loss": 0.4439, | |
| "step": 2874 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.879798903107861, | |
| "eval_loss": 0.4700252115726471, | |
| "eval_runtime": 10.7719, | |
| "eval_samples_per_second": 203.121, | |
| "eval_steps_per_second": 1.671, | |
| "step": 2874 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.4363, | |
| "step": 3832 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8761425959780622, | |
| "eval_loss": 0.4729701280593872, | |
| "eval_runtime": 10.7593, | |
| "eval_samples_per_second": 203.36, | |
| "eval_steps_per_second": 1.673, | |
| "step": 3832 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3832, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 3910307120736.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 7.264974311161276e-05, | |
| "num_train_epochs": 4, | |
| "per_device_train_batch_size": 8, | |
| "seed": 36 | |
| } | |
| } | |