trainer_state.json

{
  "best_metric": 0.9264,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0041/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": 2.2545045045045048e-05,
      "loss": 0.8396,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8866666666666667,
      "eval_loss": 0.36956194043159485,
      "eval_runtime": 6.6464,
      "eval_samples_per_second": 564.215,
      "eval_steps_per_second": 8.877,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7935595512390137,
      "learning_rate": 1.5045045045045047e-05,
      "loss": 0.1738,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9109333333333334,
      "eval_loss": 0.2938554883003235,
      "eval_runtime": 7.2348,
      "eval_samples_per_second": 518.326,
      "eval_steps_per_second": 8.155,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 8.798977851867676,
      "learning_rate": 7.545045045045046e-06,
      "loss": 0.0592,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9221333333333334,
      "eval_loss": 0.2815791070461273,
      "eval_runtime": 6.7246,
      "eval_samples_per_second": 557.656,
      "eval_steps_per_second": 8.774,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.4109489917755127,
      "learning_rate": 4.504504504504505e-08,
      "loss": 0.0166,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9264,
      "eval_loss": 0.26626238226890564,
      "eval_runtime": 6.8062,
      "eval_samples_per_second": 550.971,
      "eval_steps_per_second": 8.669,
      "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
}