trainer_state.json

{
  "best_metric": 1.0,
  "best_model_checkpoint": "vit-large-patch16-224-in21k-dungeon-geo-morphs-denoised-04Dec24-003/checkpoint-60",
  "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": 15.498607635498047,
      "learning_rate": 2e-05,
      "loss": 1.5126,
      "step": 10
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6383838383838384,
      "eval_loss": 1.33207368850708,
      "eval_runtime": 6.5954,
      "eval_samples_per_second": 75.052,
      "eval_steps_per_second": 9.4,
      "step": 10
    },
    {
      "epoch": 8.0,
      "grad_norm": 16.96432113647461,
      "learning_rate": 1.7222222222222224e-05,
      "loss": 1.049,
      "step": 20
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8444444444444444,
      "eval_loss": 0.7613236904144287,
      "eval_runtime": 6.9925,
      "eval_samples_per_second": 70.79,
      "eval_steps_per_second": 8.867,
      "step": 20
    },
    {
      "epoch": 12.0,
      "grad_norm": 18.313573837280273,
      "learning_rate": 1.4444444444444446e-05,
      "loss": 0.5397,
      "step": 30
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9434343434343434,
      "eval_loss": 0.4086134731769562,
      "eval_runtime": 6.411,
      "eval_samples_per_second": 77.211,
      "eval_steps_per_second": 9.671,
      "step": 30
    },
    {
      "epoch": 16.0,
      "grad_norm": 16.45941162109375,
      "learning_rate": 1.1666666666666668e-05,
      "loss": 0.2381,
      "step": 40
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.98989898989899,
      "eval_loss": 0.20247994363307953,
      "eval_runtime": 7.3649,
      "eval_samples_per_second": 67.21,
      "eval_steps_per_second": 8.418,
      "step": 40
    },
    {
      "epoch": 20.0,
      "grad_norm": 10.92023754119873,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.1152,
      "step": 50
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.98989898989899,
      "eval_loss": 0.11596124619245529,
      "eval_runtime": 6.844,
      "eval_samples_per_second": 72.327,
      "eval_steps_per_second": 9.059,
      "step": 50
    },
    {
      "epoch": 24.0,
      "grad_norm": 9.070709228515625,
      "learning_rate": 6.111111111111112e-06,
      "loss": 0.058,
      "step": 60
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.07251805812120438,
      "eval_runtime": 6.6661,
      "eval_samples_per_second": 74.256,
      "eval_steps_per_second": 9.301,
      "step": 60
    },
    {
      "epoch": 28.0,
      "grad_norm": 8.501797676086426,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0392,
      "step": 70
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.98989898989899,
      "eval_loss": 0.06781990826129913,
      "eval_runtime": 6.4896,
      "eval_samples_per_second": 76.276,
      "eval_steps_per_second": 9.554,
      "step": 70
    },
    {
      "epoch": 32.0,
      "grad_norm": 1.17595374584198,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.026,
      "step": 80
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.9959595959595959,
      "eval_loss": 0.04437047615647316,
      "eval_runtime": 6.7537,
      "eval_samples_per_second": 73.293,
      "eval_steps_per_second": 9.18,
      "step": 80
    },
    {
      "epoch": 32.0,
      "step": 80,
      "total_flos": 7.012786101918106e+17,
      "train_loss": 0.44722552597522736,
      "train_runtime": 549.2596,
      "train_samples_per_second": 5.826,
      "train_steps_per_second": 0.146
    }
  ],
  "logging_steps": 10,
  "max_steps": 80,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 10,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.01
      },
      "attributes": {
        "early_stopping_patience_counter": 2
      }
    },
    "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
}