{
  "id": 105345,
  "title": "How to jump from LB ~0.2 to LB ~0.5...",
  "url": "/competitions/understanding_cloud_organization/discussion/105345",
  "author_name": "",
  "post_date": "2019-08-22T14:36:37.786320900Z",
  "votes": 10,
  "comment_count": 1,
  "views": 0,
  "content": "<p>... read the description of the data. Test images don't have the same size as train images. Most people probably knows this by now, but putting this out just in case. </p>\n\n<p>Don't make the same mistake as I did :)</p>",
  "messages": [
    {
      "id": "605607",
      "postDate": "08/22/2019 14:36:37",
      "content": "<p>... read the description of the data. Test images don't have the same size as train images. Most people probably knows this by now, but putting this out just in case. </p>\n\n<p>Don't make the same mistake as I did :)</p>",
      "rawMarkdown": "... read the description of the data. Test images don't have the same size as train images. Most people probably knows this by now, but putting this out just in case. \n\nDon't make the same mistake as I did :)",
      "votes": null
    },
    {
      "id": "605635",
      "postDate": "08/22/2019 15:10:52",
      "content": "<p>Also mentioned in the first paragraph of <a href=\"https://www.kaggle.com/artgor/segmentation-in-pytorch-using-convenient-tools\">Artgor's kernel</a></p>",
      "rawMarkdown": "Also mentioned in the first paragraph of [Artgor's kernel](https://www.kaggle.com/artgor/segmentation-in-pytorch-using-convenient-tools)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 605635,
      "author_name": "xhlulu",
      "author_url": "",
      "post_date": "08/22/2019 15:10:52",
      "content": "<p>Also mentioned in the first paragraph of <a href=\"https://www.kaggle.com/artgor/segmentation-in-pytorch-using-convenient-tools\">Artgor's kernel</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "605607": "... read the description of the data. Test images don't have the same size as train images. Most people probably knows this by now, but putting this out just in case. \n\nDon't make the same mistake as I did :)",
    "605635": "Also mentioned in the first paragraph of [Artgor's kernel](https://www.kaggle.com/artgor/segmentation-in-pytorch-using-convenient-tools)"
  },
  "source": "meta"
}