{
  "id": 337417,
  "title": "Should training be done within the submission notebook?",
  "url": "/competitions/hubmap-organ-segmentation/discussion/337417",
  "author_name": "",
  "post_date": "2022-07-16T01:59:53.118592100Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>For example, the Code Requirements section of the \"HuBMAP - Hacking the Kidney\" competition states that \"for this competition training is not required in Notebooks,\" which I assume means that we are allowed to upload models we trained locally (on the given data + anything publicly available). The rules for this competition do not address this clearly, so I wanted to know what the competition hosts intended us to do.</p>",
  "messages": [
    {
      "id": "1857257",
      "postDate": "07/16/2022 01:59:53",
      "content": "<p>For example, the Code Requirements section of the \"HuBMAP - Hacking the Kidney\" competition states that \"for this competition training is not required in Notebooks,\" which I assume means that we are allowed to upload models we trained locally (on the given data + anything publicly available). The rules for this competition do not address this clearly, so I wanted to know what the competition hosts intended us to do.</p>",
      "rawMarkdown": "For example, the Code Requirements section of the \"HuBMAP - Hacking the Kidney\" competition states that \"for this competition training is not required in Notebooks,\" which I assume means that we are allowed to upload models we trained locally (on the given data + anything publicly available). The rules for this competition do not address this clearly, so I wanted to know what the competition hosts intended us to do.",
      "votes": null
    },
    {
      "id": "1857393",
      "postDate": "07/16/2022 05:16:11",
      "content": "<p>You can train models offline then upload and inference in submission notebook</p>",
      "rawMarkdown": "You can train models offline then upload and inference in submission notebook",
      "votes": null
    },
    {
      "id": "1860914",
      "postDate": "07/18/2022 16:43:45",
      "content": "<p>This is correct.</p>",
      "rawMarkdown": "This is correct.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1857393,
      "author_name": "quandapro",
      "author_url": "",
      "post_date": "07/16/2022 05:16:11",
      "content": "<p>You can train models offline then upload and inference in submission notebook</p>",
      "votes": null,
      "replies": [
        {
          "id": 1860914,
          "author_name": "sohier",
          "author_url": "",
          "post_date": "07/18/2022 16:43:45",
          "content": "<p>This is correct.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1857257": "For example, the Code Requirements section of the \"HuBMAP - Hacking the Kidney\" competition states that \"for this competition training is not required in Notebooks,\" which I assume means that we are allowed to upload models we trained locally (on the given data + anything publicly available). The rules for this competition do not address this clearly, so I wanted to know what the competition hosts intended us to do.",
    "1857393": "You can train models offline then upload and inference in submission notebook",
    "1860914": "This is correct."
  },
  "source": "meta"
}