{
  "id": 89601,
  "title": "How do you do data augmentation and data preprocess?",
  "url": "/competitions/imet-2019-fgvc6/discussion/89601",
  "author_name": "",
  "post_date": "2019-04-16T05:14:37.354041600Z",
  "votes": 6,
  "comment_count": 1,
  "views": 0,
  "content": "<p>in this <a href=\"https://www.kaggle.com/chewzy/eda-weird-images\">kernel</a>, we can see the image shapes have very large  variance. And the size and aspect ratio may be important features which related to labels. How do you do data augmentation and data preprocess? And how do you use these features？</p>",
  "messages": [
    {
      "id": "517487",
      "postDate": "04/16/2019 05:14:37",
      "content": "<p>in this <a href=\"https://www.kaggle.com/chewzy/eda-weird-images\">kernel</a>, we can see the image shapes have very large  variance. And the size and aspect ratio may be important features which related to labels. How do you do data augmentation and data preprocess? And how do you use these features？</p>",
      "rawMarkdown": "in this [kernel](https://www.kaggle.com/chewzy/eda-weird-images), we can see the image shapes have very large  variance. And the size and aspect ratio may be important features which related to labels. How do you do data augmentation and data preprocess? And how do you use these features？",
      "votes": null
    },
    {
      "id": "520069",
      "postDate": "04/20/2019 04:53:15",
      "content": "<p>image size may be a good feature to add</p>",
      "rawMarkdown": "image size may be a good feature to add",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 520069,
      "author_name": "jmourad100",
      "author_url": "",
      "post_date": "04/20/2019 04:53:15",
      "content": "<p>image size may be a good feature to add</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "517487": "in this [kernel](https://www.kaggle.com/chewzy/eda-weird-images), we can see the image shapes have very large  variance. And the size and aspect ratio may be important features which related to labels. How do you do data augmentation and data preprocess? And how do you use these features？",
    "520069": "image size may be a good feature to add"
  },
  "source": "meta"
}