| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 4460, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.7520478367805481, | |
| "learning_rate": 0.0008, | |
| "loss": 0.7704, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.5246194005012512, | |
| "eval_rouge1": 0.7067, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7048, | |
| "eval_rougeLsum": 0.7049, | |
| "eval_runtime": 39.0634, | |
| "eval_samples_per_second": 19.2, | |
| "eval_steps_per_second": 0.614, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.0620325803756714, | |
| "learning_rate": 0.0006, | |
| "loss": 0.6021, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.5193427205085754, | |
| "eval_rouge1": 0.6627, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6591, | |
| "eval_rougeLsum": 0.6597, | |
| "eval_runtime": 36.7312, | |
| "eval_samples_per_second": 20.419, | |
| "eval_steps_per_second": 0.653, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 1.2514209747314453, | |
| "learning_rate": 0.0004, | |
| "loss": 0.5536, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.4899960458278656, | |
| "eval_rouge1": 0.6414, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6433, | |
| "eval_rougeLsum": 0.641, | |
| "eval_runtime": 37.2153, | |
| "eval_samples_per_second": 20.153, | |
| "eval_steps_per_second": 0.645, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.7858202457427979, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5164, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.48993659019470215, | |
| "eval_rouge1": 0.6552, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6537, | |
| "eval_rougeLsum": 0.6541, | |
| "eval_runtime": 37.393, | |
| "eval_samples_per_second": 20.057, | |
| "eval_steps_per_second": 0.642, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.9161471724510193, | |
| "learning_rate": 0.0, | |
| "loss": 0.4856, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.4844454228878021, | |
| "eval_rouge1": 0.6575, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6536, | |
| "eval_rougeLsum": 0.6567, | |
| "eval_runtime": 37.2612, | |
| "eval_samples_per_second": 20.128, | |
| "eval_steps_per_second": 0.644, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.92219651391488e+16, | |
| "train_loss": 0.5856183791908982, | |
| "train_runtime": 3776.5703, | |
| "train_samples_per_second": 18.882, | |
| "train_steps_per_second": 1.181 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.92219651391488e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |