| { | |
| "best_metric": 0.9546666666666667, | |
| "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0525/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": 0.5578935146331787, | |
| "learning_rate": 4.665063509461097e-05, | |
| "loss": 0.7973, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9290666666666667, | |
| "eval_loss": 0.2616887092590332, | |
| "eval_runtime": 11.4385, | |
| "eval_samples_per_second": 327.839, | |
| "eval_steps_per_second": 5.158, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.11327195912599564, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 0.1217, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.944, | |
| "eval_loss": 0.19269075989723206, | |
| "eval_runtime": 11.0245, | |
| "eval_samples_per_second": 340.151, | |
| "eval_steps_per_second": 5.352, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.04619864001870155, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.0519, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9466666666666667, | |
| "eval_loss": 0.19236022233963013, | |
| "eval_runtime": 11.147, | |
| "eval_samples_per_second": 336.415, | |
| "eval_steps_per_second": 5.293, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.08641790598630905, | |
| "learning_rate": 1.2500000000000006e-05, | |
| "loss": 0.0199, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9525333333333333, | |
| "eval_loss": 0.181117445230484, | |
| "eval_runtime": 11.3402, | |
| "eval_samples_per_second": 330.681, | |
| "eval_steps_per_second": 5.203, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.05130786448717117, | |
| "learning_rate": 3.3493649053890326e-06, | |
| "loss": 0.0093, | |
| "step": 1665 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9546666666666667, | |
| "eval_loss": 0.17738869786262512, | |
| "eval_runtime": 11.3879, | |
| "eval_samples_per_second": 329.297, | |
| "eval_steps_per_second": 5.181, | |
| "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 | |
| } | |