trainer_state.json

{
  "best_metric": 0.9696969696969697,
  "best_model_checkpoint": "vit-large-patch16-224-in21k-dungeon-geo-morphs-denoised-04Dec24-001/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": 16.180925369262695,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 1.3958,
      "step": 10
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8383838383838383,
      "eval_loss": 0.9618854522705078,
      "eval_runtime": 6.2572,
      "eval_samples_per_second": 79.109,
      "eval_steps_per_second": 9.909,
      "step": 10
    },
    {
      "epoch": 8.0,
      "grad_norm": 6.994846820831299,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.4596,
      "step": 20
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9333333333333333,
      "eval_loss": 0.3876033425331116,
      "eval_runtime": 6.6276,
      "eval_samples_per_second": 74.687,
      "eval_steps_per_second": 9.355,
      "step": 20
    },
    {
      "epoch": 12.0,
      "grad_norm": 2.357105016708374,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1012,
      "step": 30
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9616161616161616,
      "eval_loss": 0.1927802562713623,
      "eval_runtime": 6.8814,
      "eval_samples_per_second": 71.933,
      "eval_steps_per_second": 9.01,
      "step": 30
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.5583979487419128,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 0.022,
      "step": 40
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9636363636363636,
      "eval_loss": 0.11810477077960968,
      "eval_runtime": 6.7616,
      "eval_samples_per_second": 73.207,
      "eval_steps_per_second": 9.169,
      "step": 40
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.18477365374565125,
      "learning_rate": 8.333333333333334e-06,
      "loss": 0.0066,
      "step": 50
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9676767676767677,
      "eval_loss": 0.09357820451259613,
      "eval_runtime": 6.4967,
      "eval_samples_per_second": 76.192,
      "eval_steps_per_second": 9.543,
      "step": 50
    },
    {
      "epoch": 24.0,
      "grad_norm": 0.11407925933599472,
      "learning_rate": 5.555555555555557e-06,
      "loss": 0.0036,
      "step": 60
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.9696969696969697,
      "eval_loss": 0.0862530842423439,
      "eval_runtime": 6.5373,
      "eval_samples_per_second": 75.72,
      "eval_steps_per_second": 9.484,
      "step": 60
    },
    {
      "epoch": 28.0,
      "grad_norm": 0.09305387735366821,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.0028,
      "step": 70
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.9696969696969697,
      "eval_loss": 0.08481135219335556,
      "eval_runtime": 7.1926,
      "eval_samples_per_second": 68.821,
      "eval_steps_per_second": 8.62,
      "step": 70
    },
    {
      "epoch": 32.0,
      "grad_norm": 0.08653579652309418,
      "learning_rate": 0.0,
      "loss": 0.0025,
      "step": 80
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.9696969696969697,
      "eval_loss": 0.08255013823509216,
      "eval_runtime": 6.77,
      "eval_samples_per_second": 73.117,
      "eval_steps_per_second": 9.158,
      "step": 80
    },
    {
      "epoch": 32.0,
      "step": 80,
      "total_flos": 7.012786101918106e+17,
      "train_loss": 0.24926918111741542,
      "train_runtime": 387.0012,
      "train_samples_per_second": 8.269,
      "train_steps_per_second": 0.207
    }
  ],
  "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
}