| { | |
| "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": 0.8314327597618103, | |
| "learning_rate": 0.0008, | |
| "loss": 1.2006, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.8184329867362976, | |
| "eval_rouge1": 0.6667, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6667, | |
| "eval_rougeLsum": 0.6663, | |
| "eval_runtime": 34.849, | |
| "eval_samples_per_second": 21.521, | |
| "eval_steps_per_second": 0.689, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.4473212659358978, | |
| "learning_rate": 0.0006, | |
| "loss": 0.6791, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.692352831363678, | |
| "eval_rouge1": 0.6654, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6613, | |
| "eval_rougeLsum": 0.6642, | |
| "eval_runtime": 34.2033, | |
| "eval_samples_per_second": 21.928, | |
| "eval_steps_per_second": 0.702, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.5131508111953735, | |
| "learning_rate": 0.0004, | |
| "loss": 0.4848, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.6633622050285339, | |
| "eval_rouge1": 0.7098, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7089, | |
| "eval_rougeLsum": 0.7131, | |
| "eval_runtime": 34.2893, | |
| "eval_samples_per_second": 21.873, | |
| "eval_steps_per_second": 0.7, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.2991909682750702, | |
| "learning_rate": 0.0002, | |
| "loss": 0.3381, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.6799816489219666, | |
| "eval_rouge1": 0.696, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.6977, | |
| "eval_rougeLsum": 0.6988, | |
| "eval_runtime": 33.9237, | |
| "eval_samples_per_second": 22.108, | |
| "eval_steps_per_second": 0.707, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.42183929681777954, | |
| "learning_rate": 0.0, | |
| "loss": 0.2027, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.7586135864257812, | |
| "eval_rouge1": 0.682, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.681, | |
| "eval_rougeLsum": 0.6795, | |
| "eval_runtime": 33.8342, | |
| "eval_samples_per_second": 22.167, | |
| "eval_steps_per_second": 0.709, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.886428467068928e+16, | |
| "train_loss": 0.5810612255147755, | |
| "train_runtime": 4303.8817, | |
| "train_samples_per_second": 16.58, | |
| "train_steps_per_second": 1.036 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.886428467068928e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |