trainer_state.json

{
  "best_metric": 0.8456,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0036/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": Infinity,
      "learning_rate": 2.5631588426322576e-05,
      "loss": 2.3276,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7450666666666667,
      "eval_loss": 1.152657151222229,
      "eval_runtime": 5.9627,
      "eval_samples_per_second": 628.906,
      "eval_steps_per_second": 9.895,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 17.318965911865234,
      "learning_rate": 1.5035378262848308e-05,
      "loss": 0.7924,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8213333333333334,
      "eval_loss": 0.6991308331489563,
      "eval_runtime": 5.8114,
      "eval_samples_per_second": 645.285,
      "eval_steps_per_second": 10.152,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 9.342079162597656,
      "learning_rate": 4.418443992810701e-06,
      "loss": 0.4325,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8352,
      "eval_loss": 0.5865416526794434,
      "eval_runtime": 5.771,
      "eval_samples_per_second": 649.803,
      "eval_steps_per_second": 10.224,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.526944398880005,
      "learning_rate": 4.172077409214658e-11,
      "loss": 0.2983,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8456,
      "eval_loss": 0.5664607286453247,
      "eval_runtime": 5.4153,
      "eval_samples_per_second": 692.488,
      "eval_steps_per_second": 10.895,
      "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
}