trainer_state.json

{
  "best_metric": 0.9389400921658986,
  "best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-gecko/checkpoint-105",
  "epoch": 4.827586206896552,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.46,
      "learning_rate": 0.00045238095238095237,
      "loss": 4.0904,
      "step": 10
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0004047619047619048,
      "loss": 3.4758,
      "step": 20
    },
    {
      "epoch": 0.97,
      "eval_accuracy": 0.565668202764977,
      "eval_loss": 3.0532968044281006,
      "eval_runtime": 99.5572,
      "eval_samples_per_second": 8.719,
      "eval_steps_per_second": 0.281,
      "step": 21
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.00035714285714285714,
      "loss": 2.9086,
      "step": 30
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00030952380952380956,
      "loss": 2.4494,
      "step": 40
    },
    {
      "epoch": 1.98,
      "eval_accuracy": 0.8191244239631337,
      "eval_loss": 2.0081708431243896,
      "eval_runtime": 96.1487,
      "eval_samples_per_second": 9.028,
      "eval_steps_per_second": 0.291,
      "step": 43
    },
    {
      "epoch": 2.3,
      "learning_rate": 0.0002619047619047619,
      "loss": 2.0186,
      "step": 50
    },
    {
      "epoch": 2.76,
      "learning_rate": 0.00021428571428571427,
      "loss": 1.6937,
      "step": 60
    },
    {
      "epoch": 2.99,
      "eval_accuracy": 0.880184331797235,
      "eval_loss": 1.426762342453003,
      "eval_runtime": 98.7923,
      "eval_samples_per_second": 8.786,
      "eval_steps_per_second": 0.283,
      "step": 65
    },
    {
      "epoch": 3.22,
      "learning_rate": 0.00016666666666666666,
      "loss": 1.4682,
      "step": 70
    },
    {
      "epoch": 3.68,
      "learning_rate": 0.00011904761904761905,
      "loss": 1.3144,
      "step": 80
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9182027649769585,
      "eval_loss": 1.1381995677947998,
      "eval_runtime": 104.0783,
      "eval_samples_per_second": 8.34,
      "eval_steps_per_second": 0.269,
      "step": 87
    },
    {
      "epoch": 4.14,
      "learning_rate": 7.142857142857142e-05,
      "loss": 1.1998,
      "step": 90
    },
    {
      "epoch": 4.6,
      "learning_rate": 2.380952380952381e-05,
      "loss": 1.1068,
      "step": 100
    },
    {
      "epoch": 4.83,
      "eval_accuracy": 0.9389400921658986,
      "eval_loss": 1.0130665302276611,
      "eval_runtime": 98.1827,
      "eval_samples_per_second": 8.841,
      "eval_steps_per_second": 0.285,
      "step": 105
    },
    {
      "epoch": 4.83,
      "step": 105,
      "total_flos": 1.0346740581105009e+18,
      "train_loss": 2.118878391810826,
      "train_runtime": 1863.9077,
      "train_samples_per_second": 7.404,
      "train_steps_per_second": 0.056
    }
  ],
  "logging_steps": 10,
  "max_steps": 105,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1.0346740581105009e+18,
  "trial_name": null,
  "trial_params": null
}