trainer_state.json

{
  "best_metric": 0.9168,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0634/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": 0.36416253447532654,
      "learning_rate": 4.525311529382417e-05,
      "loss": 0.7855,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8274666666666667,
      "eval_loss": 0.5507161617279053,
      "eval_runtime": 13.4188,
      "eval_samples_per_second": 279.458,
      "eval_steps_per_second": 4.397,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 3.281509432839476e-05,
      "loss": 0.2255,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8842666666666666,
      "eval_loss": 0.39267364144325256,
      "eval_runtime": 13.1625,
      "eval_samples_per_second": 284.9,
      "eval_steps_per_second": 4.482,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 27.21706771850586,
      "learning_rate": 1.736435462495525e-05,
      "loss": 0.0787,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8994666666666666,
      "eval_loss": 0.34574025869369507,
      "eval_runtime": 13.0219,
      "eval_samples_per_second": 287.977,
      "eval_steps_per_second": 4.531,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.012357085011899471,
      "learning_rate": 4.8301719287419e-06,
      "loss": 0.0195,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9168,
      "eval_loss": 0.3187023997306824,
      "eval_runtime": 13.0941,
      "eval_samples_per_second": 286.388,
      "eval_steps_per_second": 4.506,
      "step": 1332
    }
  ],
  "logging_steps": 500,
  "max_steps": 1665,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 6.58965282490368e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}