trainer_state.json

{
  "best_metric": 0.8968,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0047/checkpoints/checkpoint-1332",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 1332,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 3.6204991340637207,
      "learning_rate": 5.974873734152915e-05,
      "loss": 1.5899,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8032,
      "eval_loss": 0.7389627695083618,
      "eval_runtime": 6.7747,
      "eval_samples_per_second": 553.527,
      "eval_steps_per_second": 8.709,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 3.508254927997938e-05,
      "loss": 0.4933,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8525333333333334,
      "eval_loss": 0.510078489780426,
      "eval_runtime": 6.8404,
      "eval_samples_per_second": 548.211,
      "eval_steps_per_second": 8.625,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0401949882507324,
      "learning_rate": 1.0309702649891636e-05,
      "loss": 0.2677,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8904,
      "eval_loss": 0.37475812435150146,
      "eval_runtime": 7.5099,
      "eval_samples_per_second": 499.342,
      "eval_steps_per_second": 7.856,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 16.862565994262695,
      "learning_rate": 9.734847288167535e-11,
      "loss": 0.1586,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8968,
      "eval_loss": 0.3529554605484009,
      "eval_runtime": 6.8301,
      "eval_samples_per_second": 549.043,
      "eval_steps_per_second": 8.638,
      "step": 1332
    }
  ],
  "logging_steps": 500,
  "max_steps": 1332,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 6.58965282490368e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}