| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9919137466307277, | |
| "eval_steps": 500, | |
| "global_step": 46, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1078167115902965, | |
| "grad_norm": 0.25050440430641174, | |
| "learning_rate": 1.9422609221188208e-05, | |
| "loss": 1.8053, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.215633423180593, | |
| "grad_norm": 0.2306346893310547, | |
| "learning_rate": 1.77571129070442e-05, | |
| "loss": 1.7973, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.32345013477088946, | |
| "grad_norm": 0.2684953510761261, | |
| "learning_rate": 1.5195839500354338e-05, | |
| "loss": 1.7813, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.431266846361186, | |
| "grad_norm": 0.2725406587123871, | |
| "learning_rate": 1.2034560130526341e-05, | |
| "loss": 1.8348, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.5390835579514824, | |
| "grad_norm": 0.30252954363822937, | |
| "learning_rate": 8.638333509037535e-06, | |
| "loss": 1.7943, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.6469002695417789, | |
| "grad_norm": NaN, | |
| "learning_rate": 5.399349622688479e-06, | |
| "loss": 42.0508, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.7547169811320755, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.691640357218759e-06, | |
| "loss": 0.0, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.862533692722372, | |
| "grad_norm": NaN, | |
| "learning_rate": 8.278869849454718e-07, | |
| "loss": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.9703504043126685, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.3312308094607382e-08, | |
| "loss": 0.0, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.9919137466307277, | |
| "step": 46, | |
| "total_flos": 2.703782589549773e+16, | |
| "train_loss": 5.550413712211277, | |
| "train_runtime": 20332.1955, | |
| "train_samples_per_second": 0.036, | |
| "train_steps_per_second": 0.002 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 46, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 2.703782589549773e+16, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |