| { | |
| "best_metric": 0.7907170650361312, | |
| "best_model_checkpoint": "tinybert-TG-HS-HX-parentpretrained\\run-7\\checkpoint-788", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 788, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 9.391769458408264e-06, | |
| "loss": 0.4715, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7722345747637577, | |
| "eval_loss": 0.4253861606121063, | |
| "eval_runtime": 125.844, | |
| "eval_samples_per_second": 57.182, | |
| "eval_steps_per_second": 0.453, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 7.043827093806198e-06, | |
| "loss": 0.4326, | |
| "step": 394 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7833518621456365, | |
| "eval_loss": 0.4118412137031555, | |
| "eval_runtime": 125.7992, | |
| "eval_samples_per_second": 57.202, | |
| "eval_steps_per_second": 0.453, | |
| "step": 394 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 4.695884729204132e-06, | |
| "loss": 0.4188, | |
| "step": 591 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7887715397443024, | |
| "eval_loss": 0.40638259053230286, | |
| "eval_runtime": 125.6755, | |
| "eval_samples_per_second": 57.259, | |
| "eval_steps_per_second": 0.454, | |
| "step": 591 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 2.347942364602066e-06, | |
| "loss": 0.4135, | |
| "step": 788 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7907170650361312, | |
| "eval_loss": 0.40277761220932007, | |
| "eval_runtime": 125.8816, | |
| "eval_samples_per_second": 57.165, | |
| "eval_steps_per_second": 0.453, | |
| "step": 788 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 985, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 177030409545900.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.8077469751146494, | |
| "learning_rate": 1.173971182301033e-05, | |
| "num_train_epochs": 5, | |
| "temperature": 8 | |
| } | |
| } | |