| { | |
| "best_metric": 1.8619400262832642, | |
| "best_model_checkpoint": "saves/Gemma-2B/lora/train_2024-03-01-04-36-32/checkpoint-200", | |
| "epoch": 0.35555555555555557, | |
| "eval_steps": 100, | |
| "global_step": 200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 6.6079277992248535, | |
| "learning_rate": 4.999960939662063e-05, | |
| "loss": 3.747, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 3.2283411026000977, | |
| "learning_rate": 4.999843759868819e-05, | |
| "loss": 3.5789, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 41.573001861572266, | |
| "learning_rate": 4.999648464281934e-05, | |
| "loss": 3.1683, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 4.080965518951416, | |
| "learning_rate": 4.9993750590040575e-05, | |
| "loss": 2.8275, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 4.576275825500488, | |
| "learning_rate": 4.999023552578632e-05, | |
| "loss": 2.6758, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 18.012842178344727, | |
| "learning_rate": 4.998593955989626e-05, | |
| "loss": 2.6287, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 5.738934516906738, | |
| "learning_rate": 4.9980862826611875e-05, | |
| "loss": 2.5284, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 3.353776216506958, | |
| "learning_rate": 4.9975005484572305e-05, | |
| "loss": 2.2608, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 4.6298699378967285, | |
| "learning_rate": 4.9968367716809374e-05, | |
| "loss": 2.2475, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 50.594207763671875, | |
| "learning_rate": 4.996094973074183e-05, | |
| "loss": 2.2007, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "eval_loss": 2.126384735107422, | |
| "eval_runtime": 124.9221, | |
| "eval_samples_per_second": 8.005, | |
| "eval_steps_per_second": 2.001, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 10.225520133972168, | |
| "learning_rate": 4.995275175816891e-05, | |
| "loss": 1.9414, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 4.777626991271973, | |
| "learning_rate": 4.994377405526308e-05, | |
| "loss": 1.9729, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 6.133576393127441, | |
| "learning_rate": 4.993401690256203e-05, | |
| "loss": 2.0237, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 5.396271228790283, | |
| "learning_rate": 4.992348060495989e-05, | |
| "loss": 2.009, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 3.4974453449249268, | |
| "learning_rate": 4.991216549169776e-05, | |
| "loss": 2.032, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 12.256199836730957, | |
| "learning_rate": 4.990007191635334e-05, | |
| "loss": 1.9548, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 7.5634379386901855, | |
| "learning_rate": 4.988720025682995e-05, | |
| "loss": 1.8164, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 14.023727416992188, | |
| "learning_rate": 4.987355091534468e-05, | |
| "loss": 1.8517, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 4.622091293334961, | |
| "learning_rate": 4.985912431841584e-05, | |
| "loss": 2.0255, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 3.9935083389282227, | |
| "learning_rate": 4.9843920916849645e-05, | |
| "loss": 1.8777, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "eval_loss": 1.8619400262832642, | |
| "eval_runtime": 124.8712, | |
| "eval_samples_per_second": 8.008, | |
| "eval_steps_per_second": 2.002, | |
| "step": 200 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 5620, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 1.5694713402792346e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |