| { | |
| "best_metric": 0.9068051101949407, | |
| "best_model_checkpoint": "./resultsv3l32/output/checkpoint-294528", | |
| "epoch": 4.0, | |
| "global_step": 392704, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.681694100391135e-06, | |
| "loss": 0.5874, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 4.363388200782269e-06, | |
| "loss": 0.5086, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 4.045082301173403e-06, | |
| "loss": 0.48, | |
| "step": 75000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9011044943248333, | |
| "eval_loss": 0.47157734632492065, | |
| "eval_runtime": 213.0954, | |
| "eval_samples_per_second": 92.198, | |
| "eval_steps_per_second": 23.051, | |
| "step": 98176 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 3.7267764015645376e-06, | |
| "loss": 0.4559, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "learning_rate": 3.408470501955672e-06, | |
| "loss": 0.3309, | |
| "step": 125000 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "learning_rate": 3.0901646023468064e-06, | |
| "loss": 0.3398, | |
| "step": 150000 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 2.77185870273794e-06, | |
| "loss": 0.3382, | |
| "step": 175000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9044128874637349, | |
| "eval_loss": 0.5638315081596375, | |
| "eval_runtime": 212.5015, | |
| "eval_samples_per_second": 92.456, | |
| "eval_steps_per_second": 23.115, | |
| "step": 196352 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 2.4535528031290744e-06, | |
| "loss": 0.3082, | |
| "step": 200000 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "learning_rate": 2.135246903520209e-06, | |
| "loss": 0.1952, | |
| "step": 225000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 1.816941003911343e-06, | |
| "loss": 0.2029, | |
| "step": 250000 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 1.4986351043024774e-06, | |
| "loss": 0.2027, | |
| "step": 275000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9068051101949407, | |
| "eval_loss": 0.6681745052337646, | |
| "eval_runtime": 212.8042, | |
| "eval_samples_per_second": 92.324, | |
| "eval_steps_per_second": 23.082, | |
| "step": 294528 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "learning_rate": 1.1803292046936116e-06, | |
| "loss": 0.1771, | |
| "step": 300000 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "learning_rate": 8.620233050847459e-07, | |
| "loss": 0.1124, | |
| "step": 325000 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 5.437174054758801e-07, | |
| "loss": 0.1095, | |
| "step": 350000 | |
| }, | |
| { | |
| "epoch": 3.82, | |
| "learning_rate": 2.2541150586701436e-07, | |
| "loss": 0.1038, | |
| "step": 375000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9064997200590421, | |
| "eval_loss": 0.7638580799102783, | |
| "eval_runtime": 211.2725, | |
| "eval_samples_per_second": 92.994, | |
| "eval_steps_per_second": 23.25, | |
| "step": 392704 | |
| } | |
| ], | |
| "max_steps": 392704, | |
| "num_train_epochs": 4, | |
| "total_flos": 3.755067886948521e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |