krplt/GPT-Sponge
126
1{2 "best_metric": null,3 "best_model_checkpoint": null,4 "epoch": 4.760546642899584,5 "global_step": 2000,6 "is_hyper_param_search": false,7 "is_local_process_zero": true,8 "is_world_process_zero": true,9 "log_history": [10 {11 "epoch": 1.19,12 "learning_rate": 1.9920000000000002e-05,13 "loss": 1.8071,14 "step": 50015 },16 {17 "epoch": 2.38,18 "learning_rate": 3.9920000000000004e-05,19 "loss": 1.2711,20 "step": 100021 },22 {23 "epoch": 3.57,24 "learning_rate": 3.951414634146342e-05,25 "loss": 0.9311,26 "step": 150027 },28 {29 "epoch": 4.76,30 "learning_rate": 3.9026341463414634e-05,31 "loss": 0.6448,32 "step": 200033 }34 ],35 "max_steps": 42000,36 "num_train_epochs": 100,37 "total_flos": 1.4237038007161651e+17,38 "trial_name": null,39 "trial_params": null40}41 