{
  "id": 237573,
  "title": "OMG! same codes, but different score!",
  "url": "/competitions/hpa-single-cell-image-classification/discussion/237573",
  "author_name": "text2image",
  "post_date": "2021-05-09T12:38:36.325000",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>check this image：</p>\n<p>![<a href=\"http://m.qpic.cn/psc?/518322648/TmEUgtj9EK6.7V8ajmQrEOm0iAuEJfps5bptRoWriEU6pMeSRVhclcUzAQ4.P.TVt5K*MefPfXgB68kxa4gwYbRTN4MgWImx1IEZs7daqw8!/b&amp;bo=pQMMAQAAAAADF5k!&amp;rf=viewer_4&amp;t=5\" target=\"_blank\">http://m.qpic.cn/psc?/518322648/TmEUgtj9EK6.7V8ajmQrEOm0iAuEJfps5bptRoWriEU6pMeSRVhclcUzAQ4.P.TVt5K*MefPfXgB68kxa4gwYbRTN4MgWImx1IEZs7daqw8!/b&amp;bo=pQMMAQAAAAADF5k!&amp;rf=viewer_4&amp;t=5</a>]</p>\n<p>i submited the same code, but got different LB score!</p>",
  "messages": [
    {
      "id": 1299046,
      "postDate": "2021-05-09T12:38:36.327Z",
      "content": "<p>check this image：</p>\n<p>![<a href=\"http://m.qpic.cn/psc?/518322648/TmEUgtj9EK6.7V8ajmQrEOm0iAuEJfps5bptRoWriEU6pMeSRVhclcUzAQ4.P.TVt5K*MefPfXgB68kxa4gwYbRTN4MgWImx1IEZs7daqw8!/b&amp;bo=pQMMAQAAAAADF5k!&amp;rf=viewer_4&amp;t=5\" target=\"_blank\">http://m.qpic.cn/psc?/518322648/TmEUgtj9EK6.7V8ajmQrEOm0iAuEJfps5bptRoWriEU6pMeSRVhclcUzAQ4.P.TVt5K*MefPfXgB68kxa4gwYbRTN4MgWImx1IEZs7daqw8!/b&amp;bo=pQMMAQAAAAADF5k!&amp;rf=viewer_4&amp;t=5</a>]</p>\n<p>i submited the same code, but got different LB score!</p>",
      "rawMarkdown": "check this image：\n\n![http://m.qpic.cn/psc?/518322648/TmEUgtj9EK6.7V8ajmQrEOm0iAuEJfps5bptRoWriEU6pMeSRVhclcUzAQ4.P.TVt5K*MefPfXgB68kxa4gwYbRTN4MgWImx1IEZs7daqw8!/b&bo=pQMMAQAAAAADF5k!&rf=viewer_4&t=5]\n\ni submited the same code, but got different LB score!",
      "votes": 1
    },
    {
      "id": 1299161,
      "postDate": "2021-05-09T14:11:47.140Z",
      "content": "<p>That's strange, shouldn't happen if you're submitting the exact same predictions in both notebooks. But, it may not be unexpected if you're doing something like tta with different augmentations.</p>",
      "rawMarkdown": "That's strange, shouldn't happen if you're submitting the exact same predictions in both notebooks. But, it may not be unexpected if you're doing something like tta with different augmentations."
    },
    {
      "id": 1299123,
      "postDate": "2021-05-09T13:39:57.200Z",
      "content": "<p>Maybe you made changes to your dataset or you are using random test time augmentations without seeding?</p>",
      "rawMarkdown": "Maybe you made changes to your dataset or you are using random test time augmentations without seeding?"
    },
    {
      "id": 1299051,
      "postDate": "2021-05-09T12:44:39.047Z",
      "content": "<p>that's weird…..</p>",
      "rawMarkdown": "that's weird....."
    }
  ],
  "comments": [
    {
      "id": 1299161,
      "author_name": "novice03",
      "author_url": "",
      "post_date": "2021-05-09T14:11:47.140000",
      "content": "<p>That's strange, shouldn't happen if you're submitting the exact same predictions in both notebooks. But, it may not be unexpected if you're doing something like tta with different augmentations.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1299123,
      "author_name": "Alexander Riedel",
      "author_url": "",
      "post_date": "2021-05-09T13:39:57.200000",
      "content": "<p>Maybe you made changes to your dataset or you are using random test time augmentations without seeding?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1299051,
      "author_name": "text2image",
      "author_url": "",
      "post_date": "2021-05-09T12:44:39.047000",
      "content": "<p>that's weird…..</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1299046": "check this image：\n\n![http://m.qpic.cn/psc?/518322648/TmEUgtj9EK6.7V8ajmQrEOm0iAuEJfps5bptRoWriEU6pMeSRVhclcUzAQ4.P.TVt5K*MefPfXgB68kxa4gwYbRTN4MgWImx1IEZs7daqw8!/b&bo=pQMMAQAAAAADF5k!&rf=viewer_4&t=5]\n\ni submited the same code, but got different LB score!",
    "1299161": "That's strange, shouldn't happen if you're submitting the exact same predictions in both notebooks. But, it may not be unexpected if you're doing something like tta with different augmentations.",
    "1299123": "Maybe you made changes to your dataset or you are using random test time augmentations without seeding?",
    "1299051": "that's weird....."
  }
}