| { | |
| "best_metric": 0.8805333333333333, | |
| "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0864/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": 113.587158203125, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 2.6527, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7874666666666666, | |
| "eval_loss": 0.7874427437782288, | |
| "eval_runtime": 8.0983, | |
| "eval_samples_per_second": 463.059, | |
| "eval_steps_per_second": 7.285, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 51.341121673583984, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.5929, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8482666666666666, | |
| "eval_loss": 0.4722498059272766, | |
| "eval_runtime": 7.6314, | |
| "eval_samples_per_second": 491.394, | |
| "eval_steps_per_second": 7.731, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 271.38018798828125, | |
| "learning_rate": 5e-05, | |
| "loss": 0.3411, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8658666666666667, | |
| "eval_loss": 0.400565505027771, | |
| "eval_runtime": 7.6068, | |
| "eval_samples_per_second": 492.98, | |
| "eval_steps_per_second": 7.756, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 83.17887878417969, | |
| "learning_rate": 2.500000000000001e-05, | |
| "loss": 0.2292, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8741333333333333, | |
| "eval_loss": 0.3878174424171448, | |
| "eval_runtime": 7.5344, | |
| "eval_samples_per_second": 497.717, | |
| "eval_steps_per_second": 7.831, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 60.44929885864258, | |
| "learning_rate": 6.698729810778065e-06, | |
| "loss": 0.1787, | |
| "step": 1665 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8805333333333333, | |
| "eval_loss": 0.3799460232257843, | |
| "eval_runtime": 7.5205, | |
| "eval_samples_per_second": 498.637, | |
| "eval_steps_per_second": 7.845, | |
| "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": 4.088214619776e+18, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |