| { | |
| "best_metric": 0.9464, | |
| "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0077/checkpoints/checkpoint-1665", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 1665, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 3.38126540184021, | |
| "learning_rate": 4.166666666666667e-05, | |
| "loss": 0.8781, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.928, | |
| "eval_loss": 0.2700524926185608, | |
| "eval_runtime": 11.3224, | |
| "eval_samples_per_second": 331.202, | |
| "eval_steps_per_second": 5.211, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.19701194763183594, | |
| "learning_rate": 3.3333333333333335e-05, | |
| "loss": 0.1437, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9389333333333333, | |
| "eval_loss": 0.21586045622825623, | |
| "eval_runtime": 11.026, | |
| "eval_samples_per_second": 340.104, | |
| "eval_steps_per_second": 5.351, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.18565644323825836, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.0599, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.94, | |
| "eval_loss": 0.205410897731781, | |
| "eval_runtime": 11.0596, | |
| "eval_samples_per_second": 339.073, | |
| "eval_steps_per_second": 5.335, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.03939766436815262, | |
| "learning_rate": 1.6666666666666667e-05, | |
| "loss": 0.0304, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9442666666666667, | |
| "eval_loss": 0.19997914135456085, | |
| "eval_runtime": 12.2513, | |
| "eval_samples_per_second": 306.09, | |
| "eval_steps_per_second": 4.816, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.12790803611278534, | |
| "learning_rate": 8.333333333333334e-06, | |
| "loss": 0.0119, | |
| "step": 1665 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9464, | |
| "eval_loss": 0.20376107096672058, | |
| "eval_runtime": 13.3403, | |
| "eval_samples_per_second": 281.103, | |
| "eval_steps_per_second": 4.423, | |
| "step": 1665 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1998, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 6, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8.2370660311296e+18, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |