| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.0, | |
| "global_step": 10016, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 5e-05, | |
| "loss": 6.3385, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.7372845733501474e-05, | |
| "loss": 4.5084, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 4.4745691467002945e-05, | |
| "loss": 3.8034, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 4.2118537200504416e-05, | |
| "loss": 3.636, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 3.949138293400589e-05, | |
| "loss": 3.6844, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_bleu1_score": 0.1177, | |
| "eval_bleu2_score": 0.0596, | |
| "eval_bleu3_score": 0.0337, | |
| "eval_bleu4_score": 0.0194, | |
| "eval_loss": 3.402714967727661, | |
| "eval_rougeL": 0.1063, | |
| "eval_runtime": 336.0082, | |
| "eval_samples_per_second": 4.97, | |
| "eval_steps_per_second": 0.83, | |
| "step": 2504 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "learning_rate": 3.686422866750736e-05, | |
| "loss": 2.8302, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 3.423707440100883e-05, | |
| "loss": 2.8204, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "learning_rate": 3.16099201345103e-05, | |
| "loss": 2.8368, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 2.898276586801177e-05, | |
| "loss": 2.8155, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 2.635561160151324e-05, | |
| "loss": 2.8006, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_bleu1_score": 0.1478, | |
| "eval_bleu2_score": 0.078, | |
| "eval_bleu3_score": 0.046, | |
| "eval_bleu4_score": 0.028, | |
| "eval_loss": 3.2086522579193115, | |
| "eval_rougeL": 0.1402, | |
| "eval_runtime": 332.4959, | |
| "eval_samples_per_second": 5.023, | |
| "eval_steps_per_second": 0.839, | |
| "step": 5008 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 2.3728457335014714e-05, | |
| "loss": 1.888, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "learning_rate": 2.1101303068516186e-05, | |
| "loss": 1.9007, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 1.8474148802017654e-05, | |
| "loss": 1.8907, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 1.5846994535519128e-05, | |
| "loss": 1.9067, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.3219840269020598e-05, | |
| "loss": 1.8821, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_bleu1_score": 0.1664, | |
| "eval_bleu2_score": 0.0973, | |
| "eval_bleu3_score": 0.0656, | |
| "eval_bleu4_score": 0.0475, | |
| "eval_loss": 3.304532289505005, | |
| "eval_rougeL": 0.1665, | |
| "eval_runtime": 331.5903, | |
| "eval_samples_per_second": 5.036, | |
| "eval_steps_per_second": 0.841, | |
| "step": 7512 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "learning_rate": 1.0592686002522069e-05, | |
| "loss": 1.2785, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "learning_rate": 7.965531736023538e-06, | |
| "loss": 1.218, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 3.59, | |
| "learning_rate": 5.33837746952501e-06, | |
| "loss": 1.2186, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "learning_rate": 2.711223203026482e-06, | |
| "loss": 1.2381, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 3.99, | |
| "learning_rate": 8.406893652795293e-08, | |
| "loss": 1.2172, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_bleu1_score": 0.1747, | |
| "eval_bleu2_score": 0.1061, | |
| "eval_bleu3_score": 0.0748, | |
| "eval_bleu4_score": 0.0579, | |
| "eval_loss": 3.398721694946289, | |
| "eval_rougeL": 0.1708, | |
| "eval_runtime": 334.1657, | |
| "eval_samples_per_second": 4.998, | |
| "eval_steps_per_second": 0.835, | |
| "step": 10016 | |
| } | |
| ], | |
| "max_steps": 10016, | |
| "num_train_epochs": 4, | |
| "total_flos": 1.30102022111232e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |