| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.07727975270479134, | |
| "global_step": 250, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 9.949107209404665e-05, | |
| "loss": 1.6279, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 9.62558321769342e-05, | |
| "loss": 0.9748, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 8.98317278228618e-05, | |
| "loss": 0.8434, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 8.274303669726426e-05, | |
| "loss": 0.8162, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 7.19444824755478e-05, | |
| "loss": 0.7719, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 5.967492612770999e-05, | |
| "loss": 0.7687, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 4.675683150061285e-05, | |
| "loss": 0.7753, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 3.4056135837619074e-05, | |
| "loss": 0.7509, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.2424203500786474e-05, | |
| "loss": 0.75, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1.2640756555442684e-05, | |
| "loss": 0.7337, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 5.361607745106817e-06, | |
| "loss": 0.7484, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 1.0746994666109234e-06, | |
| "loss": 0.7386, | |
| "step": 240 | |
| } | |
| ], | |
| "max_steps": 250, | |
| "num_train_epochs": 1, | |
| "total_flos": 3.752851705482445e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |