{
  "id": 653656,
  "title": "DeepRC usage failure",
  "url": "/competitions/adaptive-immune-profiling-challenge-2025/discussion/653656",
  "author_name": "",
  "post_date": "2025-12-07T01:00:24.967423200Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Has anyone here used DeepRC ? I tried to use it from the library but it failed to converge even on a small dataset and the loss hovers at 0.69. For anybody who uses it and knows the reason why, any idea and advices shall be appreciated. Thank you</p>",
  "messages": [
    {
      "id": "3365119",
      "postDate": "12/07/2025 01:00:24",
      "content": "<p>Has anyone here used DeepRC ? I tried to use it from the library but it failed to converge even on a small dataset and the loss hovers at 0.69. For anybody who uses it and knows the reason why, any idea and advices shall be appreciated. Thank you</p>",
      "rawMarkdown": "Has anyone here used DeepRC ? I tried to use it from the library but it failed to converge even on a small dataset and the loss hovers at 0.69. For anybody who uses it and knows the reason why, any idea and advices shall be appreciated. Thank you",
      "votes": null
    },
    {
      "id": "3372537",
      "postDate": "12/12/2025 04:24:12",
      "content": "<p>I have the same problem</p>",
      "rawMarkdown": "I have the same problem",
      "votes": null
    },
    {
      "id": "3376253",
      "postDate": "12/14/2025 02:34:42",
      "content": "<p>See <a href=\"https://github.com/ml-jku/DeepRC/issues/9\" target=\"_blank\">https://github.com/ml-jku/DeepRC/issues/9</a> - their provided example dataset is random. I haven't seen great performance of the method on the challenge data, possibly because it leverages frequency of clones. Still trying to play around with it. </p>",
      "rawMarkdown": "See https://github.com/ml-jku/DeepRC/issues/9 - their provided example dataset is random. I haven't seen great performance of the method on the challenge data, possibly because it leverages frequency of clones. Still trying to play around with it.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3372537,
      "author_name": "micheal1994",
      "author_url": "",
      "post_date": "12/12/2025 04:24:12",
      "content": "<p>I have the same problem</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3376253,
      "author_name": "abuendia",
      "author_url": "",
      "post_date": "12/14/2025 02:34:42",
      "content": "<p>See <a href=\"https://github.com/ml-jku/DeepRC/issues/9\" target=\"_blank\">https://github.com/ml-jku/DeepRC/issues/9</a> - their provided example dataset is random. I haven't seen great performance of the method on the challenge data, possibly because it leverages frequency of clones. Still trying to play around with it. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3365119": "Has anyone here used DeepRC ? I tried to use it from the library but it failed to converge even on a small dataset and the loss hovers at 0.69. For anybody who uses it and knows the reason why, any idea and advices shall be appreciated. Thank you",
    "3372537": "I have the same problem",
    "3376253": "See https://github.com/ml-jku/DeepRC/issues/9 - their provided example dataset is random. I haven't seen great performance of the method on the challenge data, possibly because it leverages frequency of clones. Still trying to play around with it."
  },
  "source": "meta"
}