{
  "fetched_at": "2026-10-19T07:56:59Z",
  "source_url": "https://www.kaggle.com/competitions/datalab-cup3-reverse-image-caption-2021/data",
  "pages": [
    {
      "name": "data-description",
      "content": "#102flowers\n- Training images\n\n#Dataset\n- text2ImgData.pkl: a dataframe with attribute 'Captions' and 'ImagePath'\n - 'Captions' : a list of text id list contain 1 to 10 captions\n - 'ImagePath' : image path that store paired image\n- testData.pkl : a dataframe with attribute 'ID' and 'Captions'\n - 'ID': text ID used to named generated image\n - 'Captions': text id list used as condition to generate image\n\n#Dictionary\n- vocab.npy : a numpy array with all vocabulary\n- word2Id.npy : a numpy array that map word to id\n- id2Word.npy : a numpy array that map id back to word\n\n#Testing (Please do not access data in this folder)\nRunning inception_score.py requires to use GPU resource, please make sure GPU is available before run testing",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "In this work, we are interested in translating text in the form of single-sentence human-written descriptions directly into image pixels. For example, \"this flower has petals that are yellow and has a ruffled stamen\" and \"this pink and yellow flower has a beautiful yellow center with many stamens\". You have to develop a novel deep architecture and GAN formulation to effectively translate visual concepts from characters to pixels.\n\nMore specifically, given a set of texts, your task is to generate suitable images with size 64x64x3 to illustrate each of the texts.\n\nYou will compete on the modified release of [Oxford-102 flower dataset](https://www.robots.ox.ac.uk/~vgg/data/flowers/102/) and its [paired texts.](https://drive.google.com/file/d/0B0ywwgffWnLLcms2WWJQRFNSWXM/view?resourcekey=0-Av8zFbeDDvNcF1sSjDR32w)\n![image1](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2380849%2F4b114306156e8884bed87288c7bf1bd9%2Fexample.png?generation=1575453398734941&alt=media)",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "In this competition, we use both inception score and cosine distance as our final score to evaluate quality and diversity of generated images. You can find more details about inception score in this article. The final score is based on:\n\n- Similarity of images and the given contents. How similar are the generated images and the given texts?\n- KL divergence of generated images. Are the generated images very diverse?\nThe ranking on the public leaderboard is about 50% of the testing packs. Your final score(private leaderboard) will be decided by the rest of 50% after the end of the competition.\n\nThere are two baseline results, Benchmark-60 and Benchmark-80. You have to outperform Benchmark-60 to get 60 points, and Benchmark-80 to get 80. Meanwhile, the lower point(error) you achieve, the higher the final score you will get.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "This is a page where you can include rules that participants must accept before joining. You may wish to include rules like:\n\n - Don't cheat!\n - Apply yourself!\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}