| { | |
| "best_metric": 0.6358020901679993, | |
| "best_model_checkpoint": "llama_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": 2.3639440536499023, | |
| "eval_pearson": -0.09668018953689142, | |
| "eval_runtime": 15.3105, | |
| "eval_samples_per_second": 60.873, | |
| "eval_steps_per_second": 7.642, | |
| "step": 206 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.1564923524856567, | |
| "eval_pearson": 0.5814321081427005, | |
| "eval_runtime": 15.1479, | |
| "eval_samples_per_second": 61.527, | |
| "eval_steps_per_second": 7.724, | |
| "step": 412 | |
| }, | |
| { | |
| "epoch": 2.4271844660194173, | |
| "grad_norm": 20.43543243408203, | |
| "learning_rate": 0.0005145631067961165, | |
| "loss": 1.1834, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.6500518321990967, | |
| "eval_pearson": 0.8695158330097236, | |
| "eval_runtime": 15.1213, | |
| "eval_samples_per_second": 61.635, | |
| "eval_steps_per_second": 7.737, | |
| "step": 618 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.624398112297058, | |
| "eval_pearson": 0.7826281934306444, | |
| "eval_runtime": 15.343, | |
| "eval_samples_per_second": 60.744, | |
| "eval_steps_per_second": 7.626, | |
| "step": 824 | |
| }, | |
| { | |
| "epoch": 4.854368932038835, | |
| "grad_norm": 15.307787895202637, | |
| "learning_rate": 2.912621359223301e-05, | |
| "loss": 0.8975, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.6358020901679993, | |
| "eval_pearson": 0.8408963172405144, | |
| "eval_runtime": 16.7357, | |
| "eval_samples_per_second": 55.689, | |
| "eval_steps_per_second": 6.991, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 1030, | |
| "total_flos": 5.610584124311962e+16, | |
| "train_loss": 1.0273165378755735, | |
| "train_runtime": 1063.4558, | |
| "train_samples_per_second": 30.866, | |
| "train_steps_per_second": 0.969 | |
| } | |
| ], | |
| "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": 5.610584124311962e+16, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |