trainer_state.json

{
  "best_metric": 0.9130666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0045/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "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": 0.0004446112779446113,
      "loss": 0.8001,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7952,
      "eval_loss": 0.7149527072906494,
      "eval_runtime": 13.266,
      "eval_samples_per_second": 282.677,
      "eval_steps_per_second": 4.447,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 23.752063751220703,
      "learning_rate": 0.0003890557223890557,
      "loss": 0.4328,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8477333333333333,
      "eval_loss": 0.5401880741119385,
      "eval_runtime": 13.2714,
      "eval_samples_per_second": 282.562,
      "eval_steps_per_second": 4.446,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.02514801360666752,
      "learning_rate": 0.00033350016683350017,
      "loss": 0.266,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8514666666666667,
      "eval_loss": 0.5607848763465881,
      "eval_runtime": 13.2217,
      "eval_samples_per_second": 283.624,
      "eval_steps_per_second": 4.462,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": Infinity,
      "learning_rate": 0.00027811144477811147,
      "loss": 0.1787,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8637333333333334,
      "eval_loss": 0.49877291917800903,
      "eval_runtime": 13.3985,
      "eval_samples_per_second": 279.883,
      "eval_steps_per_second": 4.403,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.03814086690545082,
      "learning_rate": 0.0002225558892225559,
      "loss": 0.1077,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8784,
      "eval_loss": 0.4918215572834015,
      "eval_runtime": 13.2976,
      "eval_samples_per_second": 282.005,
      "eval_steps_per_second": 4.437,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.07750096917152405,
      "learning_rate": 0.00016700033366700035,
      "loss": 0.0593,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8906666666666667,
      "eval_loss": 0.45453372597694397,
      "eval_runtime": 13.1165,
      "eval_samples_per_second": 285.9,
      "eval_steps_per_second": 4.498,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.0013360363664105535,
      "learning_rate": 0.00011144477811144478,
      "loss": 0.025,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8933333333333333,
      "eval_loss": 0.4550786316394806,
      "eval_runtime": 12.8853,
      "eval_samples_per_second": 291.029,
      "eval_steps_per_second": 4.579,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.01317928172647953,
      "learning_rate": 5.5889222555889224e-05,
      "loss": 0.008,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9082666666666667,
      "eval_loss": 0.4193459451198578,
      "eval_runtime": 14.7792,
      "eval_samples_per_second": 253.734,
      "eval_steps_per_second": 3.992,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.010768400505185127,
      "learning_rate": 3.33667000333667e-07,
      "loss": 0.0022,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9130666666666667,
      "eval_loss": 0.4061015546321869,
      "eval_runtime": 13.2629,
      "eval_samples_per_second": 282.743,
      "eval_steps_per_second": 4.448,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "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": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}