| { | |
| "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": 0.48672914505004883, | |
| "learning_rate": 0.0008, | |
| "loss": 0.7751, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.556481659412384, | |
| "eval_rouge1": 0.6744, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6731, | |
| "eval_rougeLsum": 0.6738, | |
| "eval_runtime": 38.0578, | |
| "eval_samples_per_second": 19.707, | |
| "eval_steps_per_second": 0.631, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.4337797462940216, | |
| "learning_rate": 0.0006, | |
| "loss": 0.5945, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.5315232872962952, | |
| "eval_rouge1": 0.6624, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6627, | |
| "eval_rougeLsum": 0.6631, | |
| "eval_runtime": 38.5047, | |
| "eval_samples_per_second": 19.478, | |
| "eval_steps_per_second": 0.623, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.5190469026565552, | |
| "learning_rate": 0.0004, | |
| "loss": 0.5498, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.5278128981590271, | |
| "eval_rouge1": 0.6681, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6682, | |
| "eval_rougeLsum": 0.6679, | |
| "eval_runtime": 38.5769, | |
| "eval_samples_per_second": 19.442, | |
| "eval_steps_per_second": 0.622, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.41587188839912415, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5192, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.5092559456825256, | |
| "eval_rouge1": 0.6755, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6729, | |
| "eval_rougeLsum": 0.6744, | |
| "eval_runtime": 37.8591, | |
| "eval_samples_per_second": 19.81, | |
| "eval_steps_per_second": 0.634, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.39830541610717773, | |
| "learning_rate": 0.0, | |
| "loss": 0.4995, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.508884608745575, | |
| "eval_rouge1": 0.6799, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6778, | |
| "eval_rougeLsum": 0.681, | |
| "eval_runtime": 37.9857, | |
| "eval_samples_per_second": 19.744, | |
| "eval_steps_per_second": 0.632, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.925218464202752e+16, | |
| "train_loss": 0.5876135240221237, | |
| "train_runtime": 3946.7671, | |
| "train_samples_per_second": 18.081, | |
| "train_steps_per_second": 1.13 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.925218464202752e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |