trainer_state.json

{
  "best_metric": 0.2846003472805023,
  "best_model_checkpoint": "./vit-base-beans/checkpoint-40",
  "epoch": 2.857142857142857,
  "eval_steps": 40,
  "global_step": 40,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.71,
      "grad_norm": 0.9989465475082397,
      "learning_rate": 0.00016428571428571428,
      "loss": 0.5177,
      "step": 10
    },
    {
      "epoch": 1.43,
      "grad_norm": 1.447810173034668,
      "learning_rate": 0.00012857142857142858,
      "loss": 0.2339,
      "step": 20
    },
    {
      "epoch": 2.14,
      "grad_norm": 2.477407693862915,
      "learning_rate": 9.285714285714286e-05,
      "loss": 0.2139,
      "step": 30
    },
    {
      "epoch": 2.86,
      "grad_norm": 0.2586057484149933,
      "learning_rate": 5.714285714285714e-05,
      "loss": 0.1312,
      "step": 40
    },
    {
      "epoch": 2.86,
      "eval_accuracy": 0.8852459016393442,
      "eval_loss": 0.2846003472805023,
      "eval_runtime": 38.8967,
      "eval_samples_per_second": 1.568,
      "eval_steps_per_second": 0.206,
      "step": 40
    }
  ],
  "logging_steps": 10,
  "max_steps": 56,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 40,
  "total_flos": 4.758008162335949e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}