{
  "id": 201155,
  "title": "Classwise probability threshold calibration?",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/201155",
  "author_name": "Phaedrus",
  "post_date": "2020-12-03T12:21:50.446000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Anybody seeing improvements using probability threshold finetuning? If so, are you using some kind of brute force way?</p>\n<p>Any ideas/thoughts are appreciated.</p>",
  "messages": [
    {
      "id": 1100867,
      "postDate": "2020-12-03T12:21:50.447Z",
      "content": "<p>Anybody seeing improvements using probability threshold finetuning? If so, are you using some kind of brute force way?</p>\n<p>Any ideas/thoughts are appreciated.</p>",
      "rawMarkdown": "Anybody seeing improvements using probability threshold finetuning? If so, are you using some kind of brute force way?\n\nAny ideas/thoughts are appreciated.",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1100867": "Anybody seeing improvements using probability threshold finetuning? If so, are you using some kind of brute force way?\n\nAny ideas/thoughts are appreciated."
  }
}