trainer_state.json

{
  "best_metric": 0.9232,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0488/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": 2.1931357383728027,
      "learning_rate": 0.00042857142857142855,
      "loss": 0.7202,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8168,
      "eval_loss": 0.6467861533164978,
      "eval_runtime": 4.1395,
      "eval_samples_per_second": 905.901,
      "eval_steps_per_second": 14.253,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.7391886711120605,
      "learning_rate": 0.00035714285714285714,
      "loss": 0.3533,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8637333333333334,
      "eval_loss": 0.45458197593688965,
      "eval_runtime": 4.0513,
      "eval_samples_per_second": 925.623,
      "eval_steps_per_second": 14.563,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.6867115497589111,
      "learning_rate": 0.0002857142857142857,
      "loss": 0.1965,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8653333333333333,
      "eval_loss": 0.49402880668640137,
      "eval_runtime": 5.1197,
      "eval_samples_per_second": 732.462,
      "eval_steps_per_second": 11.524,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.0332706980407238,
      "learning_rate": 0.00021428571428571427,
      "loss": 0.1131,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8781333333333333,
      "eval_loss": 0.4780103266239166,
      "eval_runtime": 4.2119,
      "eval_samples_per_second": 890.337,
      "eval_steps_per_second": 14.008,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.01673215627670288,
      "learning_rate": 0.00014285714285714284,
      "loss": 0.0502,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9082666666666667,
      "eval_loss": 0.3552780747413635,
      "eval_runtime": 4.5528,
      "eval_samples_per_second": 823.672,
      "eval_steps_per_second": 12.959,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.010079426690936089,
      "learning_rate": 7.142857142857142e-05,
      "loss": 0.015,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9210666666666667,
      "eval_loss": 0.32777297496795654,
      "eval_runtime": 4.0594,
      "eval_samples_per_second": 923.785,
      "eval_steps_per_second": 14.534,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 2.1450021450021452e-07,
      "loss": 0.0021,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9232,
      "eval_loss": 0.31561279296875,
      "eval_runtime": 4.0919,
      "eval_samples_per_second": 916.451,
      "eval_steps_per_second": 14.419,
      "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
}