{
  "id": 219258,
  "title": "Best strategy to eval/monitort training",
  "url": "/competitions/hpa-single-cell-image-classification/discussion/219258",
  "author_name": "Strideradu",
  "post_date": "2021-02-14T04:10:16.575000",
  "votes": 11,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Since the problem is to get masks and label for each cell, I am wondering what is the local validation strategy you are using?</p>\n<p>Just classification? Or segmentation based on label and mask from cell segmenter?</p>\n<p>How is the consistency when compared to your local validation to online socre?</p>",
  "messages": [
    {
      "id": 1199718,
      "postDate": "2021-02-14T04:10:16.577Z",
      "content": "<p>Since the problem is to get masks and label for each cell, I am wondering what is the local validation strategy you are using?</p>\n<p>Just classification? Or segmentation based on label and mask from cell segmenter?</p>\n<p>How is the consistency when compared to your local validation to online socre?</p>",
      "rawMarkdown": "Since the problem is to get masks and label for each cell, I am wondering what is the local validation strategy you are using?\n\nJust classification? Or segmentation based on label and mask from cell segmenter?\n\nHow is the consistency when compared to your local validation to online socre?\n\n",
      "votes": 11
    },
    {
      "id": 1208868,
      "postDate": "2021-02-18T14:23:25.410Z",
      "content": "<p>Such a valid question with no answers as of yet. 🤕</p>",
      "rawMarkdown": "Such a valid question with no answers as of yet. 🤕",
      "votes": 1,
      "replies": [
        {
          "id": 1209460,
          "postDate": "2021-02-18T23:00:55.157Z",
          "content": "<p>I guess maybe just use eval loss of classification</p>",
          "rawMarkdown": "I guess maybe just use eval loss of classification",
          "votes": 1
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 1208868,
      "author_name": "Arka Saha",
      "author_url": "",
      "post_date": "2021-02-18T14:23:25.410000",
      "content": "<p>Such a valid question with no answers as of yet. 🤕</p>",
      "votes": 1,
      "replies": [
        {
          "id": 1209460,
          "author_name": "Strideradu",
          "author_url": "",
          "post_date": "2021-02-18T23:00:55.157000",
          "content": "<p>I guess maybe just use eval loss of classification</p>",
          "votes": 1,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1199718": "Since the problem is to get masks and label for each cell, I am wondering what is the local validation strategy you are using?\n\nJust classification? Or segmentation based on label and mask from cell segmenter?\n\nHow is the consistency when compared to your local validation to online socre?\n\n",
    "1208868": "Such a valid question with no answers as of yet. 🤕"
  }
}