trainer_state.json

{
  "best_metric": 0.8997333333333334,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0982/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": 7.202698230743408,
      "learning_rate": 7.714285714285714e-05,
      "loss": 1.6394,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7834666666666666,
      "eval_loss": 0.757558286190033,
      "eval_runtime": 8.2471,
      "eval_samples_per_second": 454.705,
      "eval_steps_per_second": 7.154,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 6.436293436293436e-05,
      "loss": 0.5258,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8381333333333333,
      "eval_loss": 0.5568339228630066,
      "eval_runtime": 7.5715,
      "eval_samples_per_second": 495.276,
      "eval_steps_per_second": 7.792,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": Infinity,
      "learning_rate": 5.154440154440155e-05,
      "loss": 0.2887,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8525333333333334,
      "eval_loss": 0.4732987582683563,
      "eval_runtime": 7.0821,
      "eval_samples_per_second": 529.507,
      "eval_steps_per_second": 8.331,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6065800786018372,
      "learning_rate": 3.868725868725869e-05,
      "loss": 0.1624,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8746666666666667,
      "eval_loss": 0.43992480635643005,
      "eval_runtime": 7.2622,
      "eval_samples_per_second": 516.371,
      "eval_steps_per_second": 8.124,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.23636570572853088,
      "learning_rate": 2.5830115830115832e-05,
      "loss": 0.0909,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8792,
      "eval_loss": 0.4315164089202881,
      "eval_runtime": 7.1504,
      "eval_samples_per_second": 524.444,
      "eval_steps_per_second": 8.251,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.045048102736473083,
      "learning_rate": 1.2972972972972973e-05,
      "loss": 0.0452,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8941333333333333,
      "eval_loss": 0.3995671272277832,
      "eval_runtime": 7.2175,
      "eval_samples_per_second": 519.573,
      "eval_steps_per_second": 8.175,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.612099647521973,
      "learning_rate": 1.1583011583011584e-07,
      "loss": 0.0207,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8997333333333334,
      "eval_loss": 0.40013906359672546,
      "eval_runtime": 7.6806,
      "eval_samples_per_second": 488.246,
      "eval_steps_per_second": 7.682,
      "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
}