trainer_state.json

{
  "best_metric": 0.9173333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0113/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": 19.05320930480957,
      "learning_rate": 0.0004849231551964771,
      "loss": 0.7363,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8290666666666666,
      "eval_loss": 0.6094328761100769,
      "eval_runtime": 6.8612,
      "eval_samples_per_second": 546.555,
      "eval_steps_per_second": 8.599,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0004416794553834106,
      "loss": 0.4009,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8608,
      "eval_loss": 0.4776949882507324,
      "eval_runtime": 5.9227,
      "eval_samples_per_second": 633.153,
      "eval_steps_per_second": 9.962,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 34.52645492553711,
      "learning_rate": 0.00037522688315464164,
      "loss": 0.2494,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8690666666666667,
      "eval_loss": 0.500095784664154,
      "eval_runtime": 5.4844,
      "eval_samples_per_second": 683.755,
      "eval_steps_per_second": 10.758,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.02480286918580532,
      "learning_rate": 0.0002936701006654613,
      "loss": 0.1477,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8888,
      "eval_loss": 0.4642309844493866,
      "eval_runtime": 5.5357,
      "eval_samples_per_second": 677.426,
      "eval_steps_per_second": 10.658,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 11.230384826660156,
      "learning_rate": 0.00020684605953398183,
      "loss": 0.0758,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8866666666666667,
      "eval_loss": 0.4735974073410034,
      "eval_runtime": 5.5399,
      "eval_samples_per_second": 676.904,
      "eval_steps_per_second": 10.65,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.12986503541469574,
      "learning_rate": 0.00012522702050703539,
      "loss": 0.0219,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9098666666666667,
      "eval_loss": 0.3993036150932312,
      "eval_runtime": 6.025,
      "eval_samples_per_second": 622.41,
      "eval_steps_per_second": 9.793,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.0024929966311901808,
      "learning_rate": 5.8657444259997555e-05,
      "loss": 0.0041,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9149333333333334,
      "eval_loss": 0.3666897118091583,
      "eval_runtime": 5.6553,
      "eval_samples_per_second": 663.097,
      "eval_steps_per_second": 10.433,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.0022826087661087513,
      "learning_rate": 1.5166604150571307e-05,
      "loss": 0.0009,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9170666666666667,
      "eval_loss": 0.35909566283226013,
      "eval_runtime": 5.4965,
      "eval_samples_per_second": 682.249,
      "eval_steps_per_second": 10.734,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.0010945458197966218,
      "learning_rate": 1.373523937919785e-10,
      "loss": 0.0007,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9173333333333333,
      "eval_loss": 0.35913822054862976,
      "eval_runtime": 5.6546,
      "eval_samples_per_second": 663.173,
      "eval_steps_per_second": 10.434,
      "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
}