| { | |
| "best_metric": 0.8160469667318982, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-40/checkpoint-768", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 768, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 7.570804595947266, | |
| "learning_rate": 9.127982348433665e-05, | |
| "loss": 0.5774, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8033268101761253, | |
| "eval_f1": 0.8222811671087534, | |
| "eval_loss": 0.45489490032196045, | |
| "eval_mcc": 0.6209445541298213, | |
| "eval_precision": 0.75, | |
| "eval_recall": 0.9099804305283757, | |
| "eval_runtime": 67.0612, | |
| "eval_samples_per_second": 15.24, | |
| "eval_steps_per_second": 0.477, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 12.614089012145996, | |
| "learning_rate": 6.0853215656224434e-05, | |
| "loss": 0.4642, | |
| "step": 384 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8091976516634051, | |
| "eval_f1": 0.8202764976958525, | |
| "eval_loss": 0.4278040826320648, | |
| "eval_mcc": 0.6231493257080627, | |
| "eval_precision": 0.7752613240418118, | |
| "eval_recall": 0.8708414872798435, | |
| "eval_runtime": 66.8566, | |
| "eval_samples_per_second": 15.286, | |
| "eval_steps_per_second": 0.479, | |
| "step": 384 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 5.883197784423828, | |
| "learning_rate": 3.0426607828112217e-05, | |
| "loss": 0.4308, | |
| "step": 576 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.815068493150685, | |
| "eval_f1": 0.8280254777070064, | |
| "eval_loss": 0.4256427586078644, | |
| "eval_mcc": 0.6374150947947472, | |
| "eval_precision": 0.7738095238095238, | |
| "eval_recall": 0.8904109589041096, | |
| "eval_runtime": 66.4637, | |
| "eval_samples_per_second": 15.377, | |
| "eval_steps_per_second": 0.481, | |
| "step": 576 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 3.9493777751922607, | |
| "learning_rate": 0.0, | |
| "loss": 0.421, | |
| "step": 768 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8160469667318982, | |
| "eval_f1": 0.8324420677361855, | |
| "eval_loss": 0.42437195777893066, | |
| "eval_mcc": 0.6445565206113043, | |
| "eval_precision": 0.7643207855973814, | |
| "eval_recall": 0.913894324853229, | |
| "eval_runtime": 66.8546, | |
| "eval_samples_per_second": 15.287, | |
| "eval_steps_per_second": 0.479, | |
| "step": 768 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 768, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 942780789120.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.9602502972445941, | |
| "learning_rate": 0.00012170643131244887, | |
| "num_train_epochs": 4, | |
| "per_device_train_batch_size": 16, | |
| "temperature": 29 | |
| } | |
| } | |