trainer_state.json

{
  "best_metric": 0.9533333333333334,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0902/checkpoints/checkpoint-2331",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2331,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.3519858419895172,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 0.7376,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.94,
      "eval_loss": 0.2254563421010971,
      "eval_runtime": 7.6382,
      "eval_samples_per_second": 490.955,
      "eval_steps_per_second": 7.724,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 17.852867126464844,
      "learning_rate": 4.9999999999999996e-05,
      "loss": 0.1176,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9434666666666667,
      "eval_loss": 0.19487839937210083,
      "eval_runtime": 7.3618,
      "eval_samples_per_second": 509.385,
      "eval_steps_per_second": 8.014,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.06716275215148926,
      "learning_rate": 3.9999999999999996e-05,
      "loss": 0.0498,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9437333333333333,
      "eval_loss": 0.2139783501625061,
      "eval_runtime": 7.4403,
      "eval_samples_per_second": 504.013,
      "eval_steps_per_second": 7.93,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.04423917829990387,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 0.0195,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9512,
      "eval_loss": 0.20353908836841583,
      "eval_runtime": 7.7344,
      "eval_samples_per_second": 484.847,
      "eval_steps_per_second": 7.628,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.09125926345586777,
      "learning_rate": 1.9999999999999998e-05,
      "loss": 0.008,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9514666666666667,
      "eval_loss": 0.20328427851200104,
      "eval_runtime": 7.6182,
      "eval_samples_per_second": 492.243,
      "eval_steps_per_second": 7.745,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.015123364515602589,
      "learning_rate": 9.999999999999999e-06,
      "loss": 0.0035,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9530666666666666,
      "eval_loss": 0.19435109198093414,
      "eval_runtime": 7.5886,
      "eval_samples_per_second": 494.162,
      "eval_steps_per_second": 7.775,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 3.0030030030030026e-08,
      "loss": 0.0019,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9533333333333334,
      "eval_loss": 0.2013852894306183,
      "eval_runtime": 7.8102,
      "eval_samples_per_second": 480.143,
      "eval_steps_per_second": 7.554,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2331,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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": 1.153189244358144e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}