trainer_state.json

{
  "best_metric": 0.4768,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0481/checkpoints/checkpoint-1665",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 1665,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0005,
      "loss": 3.21,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.22906666666666667,
      "eval_loss": 2.872424364089966,
      "eval_runtime": 6.2266,
      "eval_samples_per_second": 602.251,
      "eval_steps_per_second": 9.475,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0005,
      "loss": 2.6443,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.3096,
      "eval_loss": 2.514268159866333,
      "eval_runtime": 6.2748,
      "eval_samples_per_second": 597.627,
      "eval_steps_per_second": 9.403,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 15.67379093170166,
      "learning_rate": 0.0005,
      "loss": 2.2746,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.34613333333333335,
      "eval_loss": 2.3270559310913086,
      "eval_runtime": 6.2573,
      "eval_samples_per_second": 599.303,
      "eval_steps_per_second": 9.429,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 11.251169204711914,
      "learning_rate": 0.0005,
      "loss": 1.983,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4056,
      "eval_loss": 2.0823137760162354,
      "eval_runtime": 7.1089,
      "eval_samples_per_second": 527.506,
      "eval_steps_per_second": 8.299,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 13.21301555633545,
      "learning_rate": 0.0005,
      "loss": 1.7361,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4768,
      "eval_loss": 1.866059422492981,
      "eval_runtime": 6.3409,
      "eval_samples_per_second": 591.395,
      "eval_steps_per_second": 9.305,
      "step": 1665
    }
  ],
  "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.2370660311296e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}