| { | |
| "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": 0.5859418511390686, | |
| "learning_rate": 0.0008, | |
| "loss": 0.7784, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.5092072486877441, | |
| "eval_rouge1": 0.716, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7135, | |
| "eval_rougeLsum": 0.7135, | |
| "eval_runtime": 37.7509, | |
| "eval_samples_per_second": 19.788, | |
| "eval_steps_per_second": 0.636, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.6026113629341125, | |
| "learning_rate": 0.0006, | |
| "loss": 0.597, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.48606443405151367, | |
| "eval_rouge1": 0.7099, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7098, | |
| "eval_rougeLsum": 0.7065, | |
| "eval_runtime": 38.4767, | |
| "eval_samples_per_second": 19.414, | |
| "eval_steps_per_second": 0.624, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.7780615091323853, | |
| "learning_rate": 0.0004, | |
| "loss": 0.5538, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.47004184126853943, | |
| "eval_rouge1": 0.7231, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7238, | |
| "eval_rougeLsum": 0.7212, | |
| "eval_runtime": 39.3033, | |
| "eval_samples_per_second": 19.006, | |
| "eval_steps_per_second": 0.611, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.5704962015151978, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5245, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.4649428427219391, | |
| "eval_rouge1": 0.7304, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7271, | |
| "eval_rougeLsum": 0.7266, | |
| "eval_runtime": 37.407, | |
| "eval_samples_per_second": 19.97, | |
| "eval_steps_per_second": 0.642, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.5992371439933777, | |
| "learning_rate": 0.0, | |
| "loss": 0.5037, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.46178075671195984, | |
| "eval_rouge1": 0.7261, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7257, | |
| "eval_rougeLsum": 0.7223, | |
| "eval_runtime": 37.0144, | |
| "eval_samples_per_second": 20.181, | |
| "eval_steps_per_second": 0.648, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.923147884691456e+16, | |
| "train_loss": 0.5914828860706278, | |
| "train_runtime": 3932.4916, | |
| "train_samples_per_second": 18.139, | |
| "train_steps_per_second": 1.134 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.923147884691456e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |