| { | |
| "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": 0.5665454268455505, | |
| "learning_rate": 0.0008, | |
| "loss": 0.8225, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.5298126339912415, | |
| "eval_rouge1": 0.7091, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7089, | |
| "eval_rougeLsum": 0.7076, | |
| "eval_runtime": 37.8957, | |
| "eval_samples_per_second": 19.712, | |
| "eval_steps_per_second": 0.633, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.6210952997207642, | |
| "learning_rate": 0.0006, | |
| "loss": 0.623, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.506367027759552, | |
| "eval_rouge1": 0.7118, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.71, | |
| "eval_rougeLsum": 0.7093, | |
| "eval_runtime": 38.3838, | |
| "eval_samples_per_second": 19.461, | |
| "eval_steps_per_second": 0.625, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.7850796580314636, | |
| "learning_rate": 0.0004, | |
| "loss": 0.5809, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.4812556803226471, | |
| "eval_rouge1": 0.7155, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7147, | |
| "eval_rougeLsum": 0.7119, | |
| "eval_runtime": 37.6845, | |
| "eval_samples_per_second": 19.822, | |
| "eval_steps_per_second": 0.637, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.6170566082000732, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5549, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.4799492061138153, | |
| "eval_rouge1": 0.7058, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7057, | |
| "eval_rougeLsum": 0.7028, | |
| "eval_runtime": 38.6292, | |
| "eval_samples_per_second": 19.338, | |
| "eval_steps_per_second": 0.621, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.6198777556419373, | |
| "learning_rate": 0.0, | |
| "loss": 0.5383, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 1.0, | |
| "eval_loss": 0.4733169972896576, | |
| "eval_rouge1": 0.7008, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.7001, | |
| "eval_rougeLsum": 0.6968, | |
| "eval_runtime": 38.7242, | |
| "eval_samples_per_second": 19.29, | |
| "eval_steps_per_second": 0.62, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.90376103985152e+16, | |
| "train_loss": 0.6239253980696469, | |
| "train_runtime": 3930.1136, | |
| "train_samples_per_second": 18.15, | |
| "train_steps_per_second": 1.135 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.90376103985152e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |