trainer_state.json

{
  "best_metric": 0.14423789083957672,
  "best_model_checkpoint": "./vit-base-beans/checkpoint-80",
  "epoch": 3.6363636363636362,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.45,
      "learning_rate": 0.00017727272727272728,
      "loss": 1.2554,
      "step": 10
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.00015454545454545454,
      "loss": 0.8433,
      "step": 20
    },
    {
      "epoch": 0.91,
      "eval_accuracy": 0.9803921568627451,
      "eval_loss": 0.505466103553772,
      "eval_runtime": 66.8244,
      "eval_samples_per_second": 1.526,
      "eval_steps_per_second": 0.195,
      "step": 20
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.0001318181818181818,
      "loss": 0.4104,
      "step": 30
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.00010909090909090909,
      "loss": 0.2222,
      "step": 40
    },
    {
      "epoch": 1.82,
      "eval_accuracy": 0.9607843137254902,
      "eval_loss": 0.18820254504680634,
      "eval_runtime": 68.8172,
      "eval_samples_per_second": 1.482,
      "eval_steps_per_second": 0.189,
      "step": 40
    },
    {
      "epoch": 2.27,
      "learning_rate": 8.636363636363637e-05,
      "loss": 0.1111,
      "step": 50
    },
    {
      "epoch": 2.73,
      "learning_rate": 6.363636363636364e-05,
      "loss": 0.0797,
      "step": 60
    },
    {
      "epoch": 2.73,
      "eval_accuracy": 0.9607843137254902,
      "eval_loss": 0.1555091142654419,
      "eval_runtime": 67.1859,
      "eval_samples_per_second": 1.518,
      "eval_steps_per_second": 0.193,
      "step": 60
    },
    {
      "epoch": 3.18,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.0633,
      "step": 70
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.0554,
      "step": 80
    },
    {
      "epoch": 3.64,
      "eval_accuracy": 0.9607843137254902,
      "eval_loss": 0.14423789083957672,
      "eval_runtime": 67.9888,
      "eval_samples_per_second": 1.5,
      "eval_steps_per_second": 0.191,
      "step": 80
    }
  ],
  "max_steps": 88,
  "num_train_epochs": 4,
  "total_flos": 9.733168363084186e+16,
  "trial_name": null,
  "trial_params": null
}