trainer_state.json

{
  "best_metric": 0.9109333333333334,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0232/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.245502471923828,
      "learning_rate": 8e-05,
      "loss": 1.7008,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8184,
      "eval_loss": 0.7120881676673889,
      "eval_runtime": 13.7062,
      "eval_samples_per_second": 273.598,
      "eval_steps_per_second": 4.305,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 7.003003003003004e-05,
      "loss": 0.5017,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8466666666666667,
      "eval_loss": 0.5149756669998169,
      "eval_runtime": 13.0432,
      "eval_samples_per_second": 287.505,
      "eval_steps_per_second": 4.523,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.23260259628295898,
      "learning_rate": 6.003003003003003e-05,
      "loss": 0.2717,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8626666666666667,
      "eval_loss": 0.4893621802330017,
      "eval_runtime": 12.1616,
      "eval_samples_per_second": 308.347,
      "eval_steps_per_second": 4.851,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": Infinity,
      "learning_rate": 5.0060060060060066e-05,
      "loss": 0.1601,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8837333333333334,
      "eval_loss": 0.4137881100177765,
      "eval_runtime": 13.1741,
      "eval_samples_per_second": 284.65,
      "eval_steps_per_second": 4.478,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.1642855852842331,
      "learning_rate": 4.006006006006007e-05,
      "loss": 0.0928,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8909333333333334,
      "eval_loss": 0.4020480513572693,
      "eval_runtime": 13.307,
      "eval_samples_per_second": 281.808,
      "eval_steps_per_second": 4.434,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.08638180792331696,
      "learning_rate": 3.0060060060060065e-05,
      "loss": 0.0493,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8986666666666666,
      "eval_loss": 0.408008337020874,
      "eval_runtime": 13.0934,
      "eval_samples_per_second": 286.403,
      "eval_steps_per_second": 4.506,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.055022869259119034,
      "learning_rate": 2.0060060060060062e-05,
      "loss": 0.0232,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9037333333333334,
      "eval_loss": 0.3975636959075928,
      "eval_runtime": 13.1311,
      "eval_samples_per_second": 285.582,
      "eval_steps_per_second": 4.493,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.19397756457328796,
      "learning_rate": 1.0060060060060061e-05,
      "loss": 0.012,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9034666666666666,
      "eval_loss": 0.39462175965309143,
      "eval_runtime": 13.2183,
      "eval_samples_per_second": 283.698,
      "eval_steps_per_second": 4.464,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.10481610149145126,
      "learning_rate": 6.006006006006006e-08,
      "loss": 0.0056,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9109333333333334,
      "eval_loss": 0.37983575463294983,
      "eval_runtime": 13.2731,
      "eval_samples_per_second": 282.526,
      "eval_steps_per_second": 4.445,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}