| { | |
| "best_metric": 1.2963534593582153, | |
| "best_model_checkpoint": "mistral_sequence_classification_0.001/checkpoint-1030", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 1030, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 3.416905641555786, | |
| "eval_pearson": 0.3066219069348225, | |
| "eval_runtime": 15.9256, | |
| "eval_samples_per_second": 58.522, | |
| "eval_steps_per_second": 7.347, | |
| "step": 206 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.2889981269836426, | |
| "eval_pearson": 0.22286808706607436, | |
| "eval_runtime": 15.8126, | |
| "eval_samples_per_second": 58.94, | |
| "eval_steps_per_second": 7.399, | |
| "step": 412 | |
| }, | |
| { | |
| "epoch": 2.4271844660194173, | |
| "grad_norm": 113.98001098632812, | |
| "learning_rate": 0.0005145631067961165, | |
| "loss": 2.9323, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 1.6689746379852295, | |
| "eval_pearson": 0.301860245775783, | |
| "eval_runtime": 15.8926, | |
| "eval_samples_per_second": 58.644, | |
| "eval_steps_per_second": 7.362, | |
| "step": 618 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.4149868488311768, | |
| "eval_pearson": 0.3351186514331333, | |
| "eval_runtime": 15.655, | |
| "eval_samples_per_second": 59.534, | |
| "eval_steps_per_second": 7.474, | |
| "step": 824 | |
| }, | |
| { | |
| "epoch": 4.854368932038835, | |
| "grad_norm": 94.90880584716797, | |
| "learning_rate": 2.912621359223301e-05, | |
| "loss": 1.6676, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 1.2963534593582153, | |
| "eval_pearson": 0.339849259267878, | |
| "eval_runtime": 15.959, | |
| "eval_samples_per_second": 58.4, | |
| "eval_steps_per_second": 7.331, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 1030, | |
| "total_flos": 6.321458643082445e+16, | |
| "train_loss": 2.2687227786165995, | |
| "train_runtime": 1145.0052, | |
| "train_samples_per_second": 28.668, | |
| "train_steps_per_second": 0.9 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1030, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 6.321458643082445e+16, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |