{
  "best_metric": 0.9169753086419753,
  "best_model_checkpoint": "/content/drive/MyDrive/graduate_design/ViT_train_pictures/output_dir-full_dataset/checkpoint-8748",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 8748,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17,
      "grad_norm": 0.8087803721427917,
      "learning_rate": 0.00018856881572930955,
      "loss": 2.2007,
      "step": 500
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.5222867131233215,
      "learning_rate": 0.00017716049382716052,
      "loss": 2.2033,
      "step": 1000
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.728941023349762,
      "learning_rate": 0.00016572930955647007,
      "loss": 2.1961,
      "step": 1500
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.4614827632904053,
      "learning_rate": 0.0001542981252857796,
      "loss": 2.198,
      "step": 2000
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.6945221424102783,
      "learning_rate": 0.00014286694101508919,
      "loss": 2.1689,
      "step": 2500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3398148148148148,
      "eval_loss": 1.592108130455017,
      "eval_runtime": 114.0279,
      "eval_samples_per_second": 28.414,
      "eval_steps_per_second": 7.104,
      "step": 2916
    },
    {
      "epoch": 1.03,
      "grad_norm": 2.6183340549468994,
      "learning_rate": 0.0001314586191129401,
      "loss": 1.7575,
      "step": 3000
    },
    {
      "epoch": 1.2,
      "grad_norm": 3.532482147216797,
      "learning_rate": 0.00012002743484224966,
      "loss": 1.2952,
      "step": 3500
    },
    {
      "epoch": 1.37,
      "grad_norm": 3.6780481338500977,
      "learning_rate": 0.00010859625057155922,
      "loss": 1.0195,
      "step": 4000
    },
    {
      "epoch": 1.54,
      "grad_norm": 5.143928050994873,
      "learning_rate": 9.716506630086878e-05,
      "loss": 0.7956,
      "step": 4500
    },
    {
      "epoch": 1.71,
      "grad_norm": 10.191193580627441,
      "learning_rate": 8.575674439871972e-05,
      "loss": 0.6654,
      "step": 5000
    },
    {
      "epoch": 1.89,
      "grad_norm": 6.73974609375,
      "learning_rate": 7.432556012802928e-05,
      "loss": 0.5282,
      "step": 5500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8296296296296296,
      "eval_loss": 0.4584006369113922,
      "eval_runtime": 108.7745,
      "eval_samples_per_second": 29.786,
      "eval_steps_per_second": 7.447,
      "step": 5832
    },
    {
      "epoch": 2.06,
      "grad_norm": 6.977649211883545,
      "learning_rate": 6.294010059442158e-05,
      "loss": 0.4526,
      "step": 6000
    },
    {
      "epoch": 2.23,
      "grad_norm": 6.755834579467773,
      "learning_rate": 5.1508916323731135e-05,
      "loss": 0.3721,
      "step": 6500
    },
    {
      "epoch": 2.4,
      "grad_norm": 1.0103096961975098,
      "learning_rate": 4.007773205304069e-05,
      "loss": 0.3228,
      "step": 7000
    },
    {
      "epoch": 2.57,
      "grad_norm": 0.03946676105260849,
      "learning_rate": 2.864654778235025e-05,
      "loss": 0.28,
      "step": 7500
    },
    {
      "epoch": 2.74,
      "grad_norm": 0.23168502748012543,
      "learning_rate": 1.721536351165981e-05,
      "loss": 0.2356,
      "step": 8000
    },
    {
      "epoch": 2.91,
      "grad_norm": 4.113630771636963,
      "learning_rate": 5.807041609510746e-06,
      "loss": 0.2224,
      "step": 8500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9169753086419753,
      "eval_loss": 0.20968997478485107,
      "eval_runtime": 114.304,
      "eval_samples_per_second": 28.345,
      "eval_steps_per_second": 7.086,
      "step": 8748
    },
    {
      "epoch": 3.0,
      "step": 8748,
      "total_flos": 6.779424557667041e+18,
      "train_loss": 1.086526616065536,
      "train_runtime": 4256.354,
      "train_samples_per_second": 20.553,
      "train_steps_per_second": 2.055
    }
  ],
  "logging_steps": 500,
  "max_steps": 8748,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 6.779424557667041e+18,
  "train_batch_size": 10,
  "trial_name": null,
  "trial_params": null
}