trainer_state.json

{
  "best_metric": 0.9554666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0650/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.46573513746261597,
      "learning_rate": 4.849231551964771e-05,
      "loss": 0.8817,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9341333333333334,
      "eval_loss": 0.2542448043823242,
      "eval_runtime": 16.1278,
      "eval_samples_per_second": 232.518,
      "eval_steps_per_second": 3.658,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 4.416794553834106e-05,
      "loss": 0.1474,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9394666666666667,
      "eval_loss": 0.2166237235069275,
      "eval_runtime": 12.9852,
      "eval_samples_per_second": 288.79,
      "eval_steps_per_second": 4.544,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 7.826171398162842,
      "learning_rate": 3.7522688315464166e-05,
      "loss": 0.0696,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9509333333333333,
      "eval_loss": 0.19319765269756317,
      "eval_runtime": 12.8981,
      "eval_samples_per_second": 290.741,
      "eval_steps_per_second": 4.574,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.03971899673342705,
      "learning_rate": 2.936701006654613e-05,
      "loss": 0.0306,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9482666666666667,
      "eval_loss": 0.20276851952075958,
      "eval_runtime": 13.1049,
      "eval_samples_per_second": 286.152,
      "eval_steps_per_second": 4.502,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.016659967601299286,
      "learning_rate": 2.0684605953398183e-05,
      "loss": 0.0148,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9485333333333333,
      "eval_loss": 0.19442978501319885,
      "eval_runtime": 12.9652,
      "eval_samples_per_second": 289.235,
      "eval_steps_per_second": 4.551,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.18600670993328094,
      "learning_rate": 1.2522702050703538e-05,
      "loss": 0.0077,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9517333333333333,
      "eval_loss": 0.19406355917453766,
      "eval_runtime": 12.678,
      "eval_samples_per_second": 295.789,
      "eval_steps_per_second": 4.654,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.02953970804810524,
      "learning_rate": 5.865744425999756e-06,
      "loss": 0.0038,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9506666666666667,
      "eval_loss": 0.19045233726501465,
      "eval_runtime": 13.1174,
      "eval_samples_per_second": 285.879,
      "eval_steps_per_second": 4.498,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.03360190987586975,
      "learning_rate": 1.5166604150571306e-06,
      "loss": 0.0028,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.952,
      "eval_loss": 0.1916361153125763,
      "eval_runtime": 13.033,
      "eval_samples_per_second": 287.732,
      "eval_steps_per_second": 4.527,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.05573350936174393,
      "learning_rate": 1.3735239379197851e-11,
      "loss": 0.0028,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9554666666666667,
      "eval_loss": 0.18405339121818542,
      "eval_runtime": 13.1083,
      "eval_samples_per_second": 286.078,
      "eval_steps_per_second": 4.501,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}