{
  "fetched_at": "2026-10-20T09:02:31Z",
  "source_url": "https://www.kaggle.com/competitions/110-2-ntut-ai-intro-compet-1/data",
  "pages": [
    {
      "name": "data-description",
      "content": "This competition use the Fashion MNIST dataset created by Zalando Research. \nhttps://www.kaggle.com/zalando-research/fashionmnist\n\nThe dataset has been uploaded to Kaggle. Add the dataset directly in your Kaggle notebook. We upload the example file of prediction results here.\n\n## File\n\n*   **pred_results.csv** - Each row contains the sample ID (0 ~ 9999) and the predicted class ID.\n\n## Columns\n\n*   `Id` -  A test image's ID\n*   `Category` - Predicted class ID\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "In this competition we will train our first neural network model to classify images in Fashion MNIST ( https://www.kaggle.com/zalando-research/fashionmnist ). \n\nPlease read the code in the example notebook under \"Code\" tab. Copy and create your own notebook, design a model, train the model and make prediction on test data. Upload the prediction results of the test data. This is a multi-class classification problem and we use accuracy as the metric.\n\nIn order to identify the team's account, please create an account name starting with team number.\n參賽小組的Kaggle帳號請用分組編號作為開頭, 順便取個有趣的隊名, 例如: 1_麥塊隊\n\n每組派一名代表參賽. 只要有參加就全組期末成績+2分. 比賽第三名全組+2.5, 第二名全組+3, 第一名全組+4分!\n\n\n## Acknowledgements\nWe thank Zalando Research for providing this dataset",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "- Dicuss with your team members and submit one result!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}