{
  "id": 515842,
  "title": "Best Approaches For this competition",
  "url": "/competitions/rsna-2024-lumbar-spine-degenerative-classification/discussion/515842",
  "author_name": "",
  "post_date": "2024-06-30T08:53:21.748004700Z",
  "votes": null,
  "comment_count": 7,
  "views": 0,
  "content": "<p>I would like it a lot to hear about the approaches that you think may work or are novel for this competition. </p>\n<p>I have thought of some approaches like using an obj det model with an LSTM.<br>\nFeel free to respond. Your response will be precious.</p>",
  "messages": [
    {
      "id": "2897094",
      "postDate": "06/30/2024 08:53:21",
      "content": "<p>I would like it a lot to hear about the approaches that you think may work or are novel for this competition. </p>\n<p>I have thought of some approaches like using an obj det model with an LSTM.<br>\nFeel free to respond. Your response will be precious.</p>",
      "rawMarkdown": "I would like it a lot to hear about the approaches that you think may work or are novel for this competition. \n\nI have thought of some approaches like using an obj det model with an LSTM.\nFeel free to respond. Your response will be precious.",
      "votes": null
    },
    {
      "id": "2897230",
      "postDate": "06/30/2024 10:46:05",
      "content": "<p>I've tried LSTM with few backbones, seems to improve a little bit.<br>\nI was thinking maybe obj det model with a classification head and a object detection head using aux loss?</p>",
      "rawMarkdown": "I've tried LSTM with few backbones, seems to improve a little bit.\nI was thinking maybe obj det model with a classification head and a object detection head using aux loss?",
      "votes": null
    },
    {
      "id": "2897398",
      "postDate": "06/30/2024 13:56:45",
      "content": "<p>Can I know how did you approach to your current solution which you have submitted? I am new to kaggle</p>",
      "rawMarkdown": "Can I know how did you approach to your current solution which you have submitted? I am new to kaggle",
      "votes": null
    },
    {
      "id": "2897507",
      "postDate": "06/30/2024 14:42:31",
      "content": "<p>Does anyone know how to get 25 instances for study_id in test set without random sampling</p>",
      "rawMarkdown": "Does anyone know how to get 25 instances for study_id in test set without random sampling",
      "votes": null
    },
    {
      "id": "2898213",
      "postDate": "07/01/2024 03:21:23",
      "content": "<p>Why can u use object detection? Label the data with box by hand?</p>",
      "rawMarkdown": "Why can u use object detection? Label the data with box by hand?",
      "votes": null
    },
    {
      "id": "2900157",
      "postDate": "07/02/2024 05:03:46",
      "content": "<p>The data is already labelled and if you want to hand label, you can but you would have to make the data freely and publically available. Ref : <a href=\"https://www.kaggle.com/competitions/rsna-2024-lumbar-spine-degenerative-classification/overview/code-requirements\" target=\"_blank\">https://www.kaggle.com/competitions/rsna-2024-lumbar-spine-degenerative-classification/overview/code-requirements</a></p>",
      "rawMarkdown": "The data is already labelled and if you want to hand label, you can but you would have to make the data freely and publically available. Ref : https://www.kaggle.com/competitions/rsna-2024-lumbar-spine-degenerative-classification/overview/code-requirements",
      "votes": null
    },
    {
      "id": "2900160",
      "postDate": "07/02/2024 05:05:18",
      "content": "<p>Yeah, just take ITK8191's notebook and use a better mode and do some post processing.</p>",
      "rawMarkdown": "Yeah, just take ITK8191's notebook and use a better mode and do some post processing.",
      "votes": null
    },
    {
      "id": "2900163",
      "postDate": "07/02/2024 05:06:19",
      "content": "<p><a href=\"https://www.kaggle.com/coderrkj\" target=\"_blank\">@coderrkj</a> I would like it a lot to hear your response</p>",
      "rawMarkdown": "coderrkj I would like it a lot to hear your response",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2897230,
      "author_name": "kfk42kfk",
      "author_url": "",
      "post_date": "06/30/2024 10:46:05",
      "content": "<p>I've tried LSTM with few backbones, seems to improve a little bit.<br>\nI was thinking maybe obj det model with a classification head and a object detection head using aux loss?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2897398,
      "author_name": "sdjdkdke",
      "author_url": "",
      "post_date": "06/30/2024 13:56:45",
      "content": "<p>Can I know how did you approach to your current solution which you have submitted? I am new to kaggle</p>",
      "votes": null,
      "replies": [
        {
          "id": 2900160,
          "author_name": "devsya",
          "author_url": "",
          "post_date": "07/02/2024 05:05:18",
          "content": "<p>Yeah, just take ITK8191's notebook and use a better mode and do some post processing.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2897507,
      "author_name": "sdjdkdke",
      "author_url": "",
      "post_date": "06/30/2024 14:42:31",
      "content": "<p>Does anyone know how to get 25 instances for study_id in test set without random sampling</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2898213,
      "author_name": "fire15",
      "author_url": "",
      "post_date": "07/01/2024 03:21:23",
      "content": "<p>Why can u use object detection? Label the data with box by hand?</p>",
      "votes": null,
      "replies": [
        {
          "id": 2900157,
          "author_name": "devsya",
          "author_url": "",
          "post_date": "07/02/2024 05:03:46",
          "content": "<p>The data is already labelled and if you want to hand label, you can but you would have to make the data freely and publically available. Ref : <a href=\"https://www.kaggle.com/competitions/rsna-2024-lumbar-spine-degenerative-classification/overview/code-requirements\" target=\"_blank\">https://www.kaggle.com/competitions/rsna-2024-lumbar-spine-degenerative-classification/overview/code-requirements</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2900163,
      "author_name": "devsya",
      "author_url": "",
      "post_date": "07/02/2024 05:06:19",
      "content": "<p><a href=\"https://www.kaggle.com/coderrkj\" target=\"_blank\">@coderrkj</a> I would like it a lot to hear your response</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2897094": "I would like it a lot to hear about the approaches that you think may work or are novel for this competition. \n\nI have thought of some approaches like using an obj det model with an LSTM.\nFeel free to respond. Your response will be precious.",
    "2897230": "I've tried LSTM with few backbones, seems to improve a little bit.\nI was thinking maybe obj det model with a classification head and a object detection head using aux loss?",
    "2897398": "Can I know how did you approach to your current solution which you have submitted? I am new to kaggle",
    "2897507": "Does anyone know how to get 25 instances for study_id in test set without random sampling",
    "2898213": "Why can u use object detection? Label the data with box by hand?",
    "2900157": "The data is already labelled and if you want to hand label, you can but you would have to make the data freely and publically available. Ref : https://www.kaggle.com/competitions/rsna-2024-lumbar-spine-degenerative-classification/overview/code-requirements",
    "2900160": "Yeah, just take ITK8191's notebook and use a better mode and do some post processing.",
    "2900163": "coderrkj I would like it a lot to hear your response"
  },
  "source": "meta"
}