trainer_state.json

{
  "best_metric": 0.9964285714285714,
  "best_model_checkpoint": "vit-large-patch16-224-dungeon-geo-morphs-0-4-29Nov24-002/checkpoint-80",
  "epoch": 32.0,
  "eval_steps": 10,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 4.0,
      "grad_norm": 30.849088668823242,
      "learning_rate": 9.861111111111112e-06,
      "loss": 1.576,
      "step": 10
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5803571428571429,
      "eval_loss": 1.2765302658081055,
      "eval_runtime": 7.5512,
      "eval_samples_per_second": 74.161,
      "eval_steps_per_second": 9.27,
      "step": 10
    },
    {
      "epoch": 8.0,
      "grad_norm": 23.055137634277344,
      "learning_rate": 8.472222222222223e-06,
      "loss": 1.0281,
      "step": 20
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8339285714285715,
      "eval_loss": 0.7735558748245239,
      "eval_runtime": 8.2438,
      "eval_samples_per_second": 67.93,
      "eval_steps_per_second": 8.491,
      "step": 20
    },
    {
      "epoch": 12.0,
      "grad_norm": 20.051189422607422,
      "learning_rate": 7.083333333333335e-06,
      "loss": 0.594,
      "step": 30
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9196428571428571,
      "eval_loss": 0.42900553345680237,
      "eval_runtime": 7.5773,
      "eval_samples_per_second": 73.905,
      "eval_steps_per_second": 9.238,
      "step": 30
    },
    {
      "epoch": 16.0,
      "grad_norm": 14.912154197692871,
      "learning_rate": 5.694444444444445e-06,
      "loss": 0.3375,
      "step": 40
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9660714285714286,
      "eval_loss": 0.264893114566803,
      "eval_runtime": 8.3101,
      "eval_samples_per_second": 67.388,
      "eval_steps_per_second": 8.424,
      "step": 40
    },
    {
      "epoch": 20.0,
      "grad_norm": 14.604040145874023,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.2094,
      "step": 50
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9857142857142858,
      "eval_loss": 0.15899096429347992,
      "eval_runtime": 8.5413,
      "eval_samples_per_second": 65.564,
      "eval_steps_per_second": 8.196,
      "step": 50
    },
    {
      "epoch": 24.0,
      "grad_norm": 11.598795890808105,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.1342,
      "step": 60
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.9928571428571429,
      "eval_loss": 0.11229041963815689,
      "eval_runtime": 7.8531,
      "eval_samples_per_second": 71.309,
      "eval_steps_per_second": 8.914,
      "step": 60
    },
    {
      "epoch": 28.0,
      "grad_norm": 7.593981742858887,
      "learning_rate": 1.527777777777778e-06,
      "loss": 0.1041,
      "step": 70
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.9928571428571429,
      "eval_loss": 0.09983398020267487,
      "eval_runtime": 8.2646,
      "eval_samples_per_second": 67.758,
      "eval_steps_per_second": 8.47,
      "step": 70
    },
    {
      "epoch": 32.0,
      "grad_norm": 8.414456367492676,
      "learning_rate": 1.3888888888888888e-07,
      "loss": 0.0832,
      "step": 80
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.9964285714285714,
      "eval_loss": 0.08847405761480331,
      "eval_runtime": 8.0748,
      "eval_samples_per_second": 69.352,
      "eval_steps_per_second": 8.669,
      "step": 80
    },
    {
      "epoch": 32.0,
      "step": 80,
      "total_flos": 7.012786101918106e+17,
      "train_loss": 0.5083139404654503,
      "train_runtime": 494.5702,
      "train_samples_per_second": 6.47,
      "train_steps_per_second": 0.162
    }
  ],
  "logging_steps": 10,
  "max_steps": 80,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 10,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.012786101918106e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}