{
  "id": 294977,
  "title": "Breaking up the multiple descriptions in each set",
  "url": "/competitions/cs5785-fall-2021-final/discussion/294977",
  "author_name": "JwallTester",
  "post_date": "2021-12-13T19:34:03.866000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>For each image we are given 5 different 1 sentence descriptions. The preprocessing of the baseline version makes these 5 descriptions a single string. However, when training my model, I have found success extending the dataset by making each sentence - resnet pair its own sample.</p>",
  "messages": [
    {
      "id": 1617028,
      "postDate": "2021-12-13T19:34:03.867Z",
      "content": "<p>For each image we are given 5 different 1 sentence descriptions. The preprocessing of the baseline version makes these 5 descriptions a single string. However, when training my model, I have found success extending the dataset by making each sentence - resnet pair its own sample.</p>",
      "rawMarkdown": "For each image we are given 5 different 1 sentence descriptions. The preprocessing of the baseline version makes these 5 descriptions a single string. However, when training my model, I have found success extending the dataset by making each sentence - resnet pair its own sample.",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1617028": "For each image we are given 5 different 1 sentence descriptions. The preprocessing of the baseline version makes these 5 descriptions a single string. However, when training my model, I have found success extending the dataset by making each sentence - resnet pair its own sample."
  }
}