| { | |
| "best_metric": 0.8455210237659964, | |
| "best_model_checkpoint": "distilbert-hate_speech18\\run-2\\checkpoint-1437", | |
| "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": 3.7733581062071546e-06, | |
| "loss": 0.1726, | |
| "step": 479 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6846435100548446, | |
| "eval_loss": 0.1708744764328003, | |
| "eval_runtime": 8.1063, | |
| "eval_samples_per_second": 269.915, | |
| "eval_steps_per_second": 16.901, | |
| "step": 479 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 2.8300185796553657e-06, | |
| "loss": 0.1708, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8244972577696527, | |
| "eval_loss": 0.17035162448883057, | |
| "eval_runtime": 7.9955, | |
| "eval_samples_per_second": 273.654, | |
| "eval_steps_per_second": 17.135, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.8866790531035773e-06, | |
| "loss": 0.1699, | |
| "step": 1437 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8455210237659964, | |
| "eval_loss": 0.17017477750778198, | |
| "eval_runtime": 8.0232, | |
| "eval_samples_per_second": 272.708, | |
| "eval_steps_per_second": 17.075, | |
| "step": 1437 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 9.433395265517887e-07, | |
| "loss": 0.1694, | |
| "step": 1916 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.829981718464351, | |
| "eval_loss": 0.17003203928470612, | |
| "eval_runtime": 8.0266, | |
| "eval_samples_per_second": 272.595, | |
| "eval_steps_per_second": 17.068, | |
| "step": 1916 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2395, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 514695596916480.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.1262637885395267, | |
| "learning_rate": 4.716697632758943e-06, | |
| "num_train_epochs": 5, | |
| "per_device_eval_batch_size": 16, | |
| "per_device_train_batch_size": 16, | |
| "temperature": 6 | |
| } | |
| } | |