{
  "id": 576462,
  "title": "What is the TM-score of the single trained Protenix  Model?",
  "url": "/competitions/stanford-rna-3d-folding/discussion/576462",
  "author_name": "",
  "post_date": "2025-05-05T10:28:52.498639300Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hello Kaggle community,</p>\n<p>I'm currently exploring the Protenix model  and would like to understand its performance more clearly. Could someone please provide information about:</p>\n<p>The TM-score achieved by the single trained Protenix model on standard  dataset</p>\n<p>How does its accuracy compare to other models </p>",
  "messages": [
    {
      "id": "3194012",
      "postDate": "05/05/2025 10:28:52",
      "content": "<p>Hello Kaggle community,</p>\n<p>I'm currently exploring the Protenix model  and would like to understand its performance more clearly. Could someone please provide information about:</p>\n<p>The TM-score achieved by the single trained Protenix model on standard  dataset</p>\n<p>How does its accuracy compare to other models </p>",
      "rawMarkdown": "Hello Kaggle community,\n\nI'm currently exploring the Protenix model  and would like to understand its performance more clearly. Could someone please provide information about:\n\nThe TM-score achieved by the single trained Protenix model on standard  dataset\n\nHow does its accuracy compare to other models",
      "votes": null
    },
    {
      "id": "3194335",
      "postDate": "05/05/2025 17:41:47",
      "content": "<p>I got a LB score of 0.355 using the protenix model out of the box if thats what ur asking</p>",
      "rawMarkdown": "I got a LB score of 0.355 using the protenix model out of the box if thats what ur asking",
      "votes": null
    },
    {
      "id": "3194673",
      "postDate": "05/06/2025 07:18:57",
      "content": "<p>From my perspective, Protenix appears to perform better than other models specifically designed for RNA structure prediction. I am curious about the reasons behind this (simply based on a comparison of pretrained models).</p>",
      "rawMarkdown": "From my perspective, Protenix appears to perform better than other models specifically designed for RNA structure prediction. I am curious about the reasons behind this (simply based on a comparison of pretrained models).",
      "votes": null
    },
    {
      "id": "3194674",
      "postDate": "05/06/2025 07:21:54",
      "content": "<p>By conducting repeated trials, you can achieve varying scores that may potentially exceed 0.355 (since the process lacks full firm reproducibility; for instance, a competitor achieved a score of 0.371 in the competition through iterative attempts). However, it is important not to be misled by the LB.</p>",
      "rawMarkdown": "By conducting repeated trials, you can achieve varying scores that may potentially exceed 0.355 (since the process lacks full firm reproducibility; for instance, a competitor achieved a score of 0.371 in the competition through iterative attempts). However, it is important not to be misled by the LB.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3194335,
      "author_name": "nikhileshbelulkar",
      "author_url": "",
      "post_date": "05/05/2025 17:41:47",
      "content": "<p>I got a LB score of 0.355 using the protenix model out of the box if thats what ur asking</p>",
      "votes": null,
      "replies": [
        {
          "id": 3194674,
          "author_name": "nigmatrahim4",
          "author_url": "",
          "post_date": "05/06/2025 07:21:54",
          "content": "<p>By conducting repeated trials, you can achieve varying scores that may potentially exceed 0.355 (since the process lacks full firm reproducibility; for instance, a competitor achieved a score of 0.371 in the competition through iterative attempts). However, it is important not to be misled by the LB.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3194673,
      "author_name": "nigmatrahim4",
      "author_url": "",
      "post_date": "05/06/2025 07:18:57",
      "content": "<p>From my perspective, Protenix appears to perform better than other models specifically designed for RNA structure prediction. I am curious about the reasons behind this (simply based on a comparison of pretrained models).</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3194012": "Hello Kaggle community,\n\nI'm currently exploring the Protenix model  and would like to understand its performance more clearly. Could someone please provide information about:\n\nThe TM-score achieved by the single trained Protenix model on standard  dataset\n\nHow does its accuracy compare to other models",
    "3194335": "I got a LB score of 0.355 using the protenix model out of the box if thats what ur asking",
    "3194673": "From my perspective, Protenix appears to perform better than other models specifically designed for RNA structure prediction. I am curious about the reasons behind this (simply based on a comparison of pretrained models).",
    "3194674": "By conducting repeated trials, you can achieve varying scores that may potentially exceed 0.355 (since the process lacks full firm reproducibility; for instance, a competitor achieved a score of 0.371 in the competition through iterative attempts). However, it is important not to be misled by the LB."
  },
  "source": "meta"
}