| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.983783783783784, | |
| "eval_steps": 500, | |
| "global_step": 276, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 257.5615234375, | |
| "learning_rate": 1.4285714285714286e-06, | |
| "loss": 9.3335, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 7.7999067306518555, | |
| "learning_rate": 1.954198473282443e-05, | |
| "loss": 1.7156, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 6.204021453857422, | |
| "learning_rate": 1.8015267175572518e-05, | |
| "loss": 0.2954, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 6.83119010925293, | |
| "learning_rate": 1.648854961832061e-05, | |
| "loss": 0.287, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 5.538864612579346, | |
| "learning_rate": 1.4961832061068704e-05, | |
| "loss": 0.2354, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 4.499451637268066, | |
| "learning_rate": 1.3435114503816796e-05, | |
| "loss": 0.1921, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 3.2935373783111572, | |
| "learning_rate": 1.1908396946564887e-05, | |
| "loss": 0.1084, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 4.249820232391357, | |
| "learning_rate": 1.0381679389312977e-05, | |
| "loss": 0.1014, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "grad_norm": 3.5348477363586426, | |
| "learning_rate": 8.85496183206107e-06, | |
| "loss": 0.1095, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 2.605787992477417, | |
| "learning_rate": 7.328244274809161e-06, | |
| "loss": 0.1038, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 2.635464668273926, | |
| "learning_rate": 5.801526717557252e-06, | |
| "loss": 0.058, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "grad_norm": 2.7133588790893555, | |
| "learning_rate": 4.274809160305344e-06, | |
| "loss": 0.0549, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "grad_norm": 2.451127767562866, | |
| "learning_rate": 2.7480916030534356e-06, | |
| "loss": 0.0526, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "grad_norm": 4.406803607940674, | |
| "learning_rate": 1.2213740458015268e-06, | |
| "loss": 0.0482, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "step": 276, | |
| "total_flos": 9.00581770395648e+18, | |
| "train_loss": 0.27409684614858765, | |
| "train_runtime": 3328.7246, | |
| "train_samples_per_second": 2.664, | |
| "train_steps_per_second": 0.083 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 276, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 300, | |
| "total_flos": 9.00581770395648e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |