{
  "id": 272446,
  "title": "my thought about this competition",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/272446",
  "author_name": "",
  "post_date": "2021-09-15T13:39:29.203576Z",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi, everyone!<br>\n    I'm new to this competition just now. I don't know whether it's too late. Anyway, I will just make a nice model.<br>\n    After reading some code, based on my knowledge about the medical images, I don't think K-fold will help a lot. I just guess that, maybe wrong. <br>\n    My idea is that the image is not enough to train a robust model just like that. I will try semi-supervise.  <br>\n    And I'd like to try the model like 3D U-net. I don't know whether it's too large to train. Any one try it?  please tell me. Thanks.</p>",
  "messages": [
    {
      "id": "1513860",
      "postDate": "09/15/2021 13:39:29",
      "content": "<p>Hi, everyone!<br>\n    I'm new to this competition just now. I don't know whether it's too late. Anyway, I will just make a nice model.<br>\n    After reading some code, based on my knowledge about the medical images, I don't think K-fold will help a lot. I just guess that, maybe wrong. <br>\n    My idea is that the image is not enough to train a robust model just like that. I will try semi-supervise.  <br>\n    And I'd like to try the model like 3D U-net. I don't know whether it's too large to train. Any one try it?  please tell me. Thanks.</p>",
      "rawMarkdown": "Hi, everyone!\n    I'm new to this competition just now. I don't know whether it's too late. Anyway, I will just make a nice model.\n    After reading some code, based on my knowledge about the medical images, I don't think K-fold will help a lot. I just guess that, maybe wrong. \n    My idea is that the image is not enough to train a robust model just like that. I will try semi-supervise.  \n    And I'd like to try the model like 3D U-net. I don't know whether it's too large to train. Any one try it?  please tell me. Thanks.",
      "votes": null
    },
    {
      "id": "1514061",
      "postDate": "09/15/2021 16:44:09",
      "content": "<p>The dataset seems way too small to train anything good :(</p>",
      "rawMarkdown": "The dataset seems way too small to train anything good :(",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1514061,
      "author_name": "zarif98sjs",
      "author_url": "",
      "post_date": "09/15/2021 16:44:09",
      "content": "<p>The dataset seems way too small to train anything good :(</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1513860": "Hi, everyone!\n    I'm new to this competition just now. I don't know whether it's too late. Anyway, I will just make a nice model.\n    After reading some code, based on my knowledge about the medical images, I don't think K-fold will help a lot. I just guess that, maybe wrong. \n    My idea is that the image is not enough to train a robust model just like that. I will try semi-supervise.  \n    And I'd like to try the model like 3D U-net. I don't know whether it's too large to train. Any one try it?  please tell me. Thanks.",
    "1514061": "The dataset seems way too small to train anything good :("
  },
  "source": "meta"
}