{
  "id": 576526,
  "title": "Nvidia co-hosts!",
  "url": "/competitions/stanford-rna-3d-folding/discussion/576526",
  "author_name": "Rhiju Das",
  "post_date": "2025-05-05T15:06:51.262000",
  "votes": 14,
  "comment_count": 9,
  "views": 0,
  "content": "<p>Hi everyone, it's great to see the rapid progress in RNA 3D structure prediction with so many participants and the sharing of novel approaches.</p>\n<p>As we enter the final month, we have some exciting news to share.</p>\n<p>As you may know, Nvidia has been helping&nbsp;@shujun17&nbsp;and our lab train the 100M RibonanzaNet2 (RNet2) model.</p>\n<p>We released the alpha version of RNet2 on Kaggle&nbsp;last month&nbsp;and the model was highlighted by Nvidia in their&nbsp;<a href=\"https://developer.nvidia.com/blog/stanford-das-lab-accelerates-rna-folding-research-with-nvidia-dgx-cloud/\" target=\"_blank\">developer blog</a>.</p>\n<p>Now, our Nvidia research collaborator&nbsp;@chrismunley&nbsp;as well as Kaggle grandmasters&nbsp;@youhanlee&nbsp;and KGMoN member&nbsp;@theoveil&nbsp;are joining as official co-hosts of this competition!</p>\n<p>You may see them providing some interesting baseline models and answering questions on forums. We are very excited to have Nvidia's scientific collaboration in tackling this grand challenge in biology.</p>",
  "messages": [
    {
      "id": 3194234,
      "postDate": "2025-05-05T15:06:51.263Z",
      "content": "<p>Hi everyone, it's great to see the rapid progress in RNA 3D structure prediction with so many participants and the sharing of novel approaches.</p>\n<p>As we enter the final month, we have some exciting news to share.</p>\n<p>As you may know, Nvidia has been helping&nbsp;@shujun17&nbsp;and our lab train the 100M RibonanzaNet2 (RNet2) model.</p>\n<p>We released the alpha version of RNet2 on Kaggle&nbsp;last month&nbsp;and the model was highlighted by Nvidia in their&nbsp;<a href=\"https://developer.nvidia.com/blog/stanford-das-lab-accelerates-rna-folding-research-with-nvidia-dgx-cloud/\" target=\"_blank\">developer blog</a>.</p>\n<p>Now, our Nvidia research collaborator&nbsp;@chrismunley&nbsp;as well as Kaggle grandmasters&nbsp;@youhanlee&nbsp;and KGMoN member&nbsp;@theoveil&nbsp;are joining as official co-hosts of this competition!</p>\n<p>You may see them providing some interesting baseline models and answering questions on forums. We are very excited to have Nvidia's scientific collaboration in tackling this grand challenge in biology.</p>",
      "rawMarkdown": "Hi everyone, it's great to see the rapid progress in RNA 3D structure prediction with so many participants and the sharing of novel approaches.\n\nAs we enter the final month, we have some exciting news to share.\n\nAs you may know, Nvidia has been helping @shujun17 and our lab train the 100M RibonanzaNet2 (RNet2) model.\n\nWe released the alpha version of RNet2 on Kaggle last month and the model was highlighted by Nvidia in their [developer blog](https://developer.nvidia.com/blog/stanford-das-lab-accelerates-rna-folding-research-with-nvidia-dgx-cloud/).\n\nNow, our Nvidia research collaborator @chrismunley as well as Kaggle grandmasters @youhanlee and KGMoN member @theoveil are joining as official co-hosts of this competition!\n\nYou may see them providing some interesting baseline models and answering questions on forums. We are very excited to have Nvidia's scientific collaboration in tackling this grand challenge in biology.\n",
      "votes": 14
    },
    {
      "id": 3194834,
      "postDate": "2025-05-06T10:55:52.953Z",
      "content": "<p><a href=\"https://www.kaggle.com/rhijudas\" target=\"_blank\">@rhijudas</a> I’m concerned GMs might release code that break the Leaderboard, specially now that we are on the last stage of the competition. I am in favor of sharing but not spoon feeding. Its not fair for the ones who have worked hard on their submissions</p>",
      "rawMarkdown": "@rhijudas I’m concerned GMs might release code that break the Leaderboard, specially now that we are on the last stage of the competition. I am in favor of sharing but not spoon feeding. Its not fair for the ones who have worked hard on their submissions",
      "votes": -1,
      "replies": [
        {
          "id": 3194880,
          "postDate": "2025-05-06T11:38:47.933Z",
          "rawMarkdown": "",
          "isDeleted": true
        },
        {
          "id": 3194881,
          "postDate": "2025-05-06T11:39:24.817Z",
          "content": "<p>I believe the hosts will provide baselines in a fair and thoughtful way that benefits all Kagglers participating in the competition!</p>",
          "rawMarkdown": "I believe the hosts will provide baselines in a fair and thoughtful way that benefits all Kagglers participating in the competition!"
        },
        {
          "id": 3194892,
          "postDate": "2025-05-06T12:00:17.973Z",
          "content": "<p>I agree with moth <a href=\"https://www.kaggle.com/alejopaullier\" target=\"_blank\">@alejopaullier</a> </p>\n<p>It took me a few weeks just to implement a novel algorithm I came up with so training can be done on long sequences (&gt; 1000) <em>without any</em> low-dimension projection sacrifice purely for computation saving.</p>\n<p>Kick in a few such \"novel algorithm\" examples, and you can guess why I haven't even submitted yet. (Soon, though, soon…)</p>\n<p>Sharing is great, but this late into the competition, maybe we should extend the deadline a little to give the hardworking teams opportunities to flex on the shared solution compared to the spoonfeeding teams.</p>\n<p>As the final RNA test will be as late as September, a bit of extension to neutralize spoon-feeding isn't too much of a deal, is it?</p>",
          "rawMarkdown": "I agree with moth @alejopaullier \n\nIt took me a few weeks just to implement a novel algorithm I came up with so training can be done on long sequences (> 1000) *without any* low-dimension projection sacrifice purely for computation saving.\n\nKick in a few such \"novel algorithm\" examples, and you can guess why I haven't even submitted yet. (Soon, though, soon...)\n\nSharing is great, but this late into the competition, maybe we should extend the deadline a little to give the hardworking teams opportunities to flex on the shared solution compared to the spoonfeeding teams.\n\nAs the final RNA test will be as late as September, a bit of extension to neutralize spoon-feeding isn't too much of a deal, is it?"
        },
        {
          "id": 3194904,
          "postDate": "2025-05-06T12:06:27.547Z",
          "content": "<p>I am also concerned about the Early Sharing Prize. I haven't seen a deadline for it. Does it mean it will be possible to make public a high scoring notebook on the last days of the competition? This would absolutely break the LB and it is completely against Kaggle spirit. Kaggle usually disallows sharing high scoring solutions after the last two weeks of any competition.</p>",
          "rawMarkdown": "I am also concerned about the Early Sharing Prize. I haven't seen a deadline for it. Does it mean it will be possible to make public a high scoring notebook on the last days of the competition? This would absolutely break the LB and it is completely against Kaggle spirit. Kaggle usually disallows sharing high scoring solutions after the last two weeks of any competition.",
          "votes": -1,
          "replies": [
            {
              "id": 3196666,
              "postDate": "2025-05-07T09:44:09.603Z",
              "content": "<blockquote>\n  <p>Kaggle usually disallows sharing high scoring solutions after the last two weeks of any competition.  </p>\n</blockquote>\n<p>Only in the last week.  </p>",
              "rawMarkdown": ">Kaggle usually disallows sharing high scoring solutions after the last two weeks of any competition.  \n\nOnly in the last week.  ",
              "votes": 1
            },
            {
              "id": 3196846,
              "postDate": "2025-05-07T13:49:03.207Z",
              "content": "<p>Thanks, I thought it was the last 2</p>",
              "rawMarkdown": "Thanks, I thought it was the last 2"
            }
          ]
        }
      ]
    },
    {
      "id": 3197211,
      "postDate": "2025-05-07T22:19:53.557Z",
      "content": "<p>its just amazing!</p>",
      "rawMarkdown": "its just amazing!"
    },
    {
      "id": 3194676,
      "postDate": "2025-05-06T07:26:15.963Z",
      "content": "<p>Encouraging news!!!</p>",
      "rawMarkdown": "Encouraging news!!!"
    }
  ],
  "comments": [
    {
      "id": 3194834,
      "author_name": "moth",
      "author_url": "",
      "post_date": "2025-05-06T10:55:52.953000",
      "content": "<p><a href=\"https://www.kaggle.com/rhijudas\" target=\"_blank\">@rhijudas</a> I’m concerned GMs might release code that break the Leaderboard, specially now that we are on the last stage of the competition. I am in favor of sharing but not spoon feeding. Its not fair for the ones who have worked hard on their submissions</p>",
      "votes": -1,
      "replies": [
        {
          "id": 3194880,
          "author_name": "",
          "author_url": "",
          "post_date": "2025-05-06T11:38:47.933000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 3194881,
          "author_name": "doheon114",
          "author_url": "",
          "post_date": "2025-05-06T11:39:24.817000",
          "content": "<p>I believe the hosts will provide baselines in a fair and thoughtful way that benefits all Kagglers participating in the competition!</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 3194892,
          "author_name": "Truth Seeker",
          "author_url": "",
          "post_date": "2025-05-06T12:00:17.973000",
          "content": "<p>I agree with moth <a href=\"https://www.kaggle.com/alejopaullier\" target=\"_blank\">@alejopaullier</a> </p>\n<p>It took me a few weeks just to implement a novel algorithm I came up with so training can be done on long sequences (&gt; 1000) <em>without any</em> low-dimension projection sacrifice purely for computation saving.</p>\n<p>Kick in a few such \"novel algorithm\" examples, and you can guess why I haven't even submitted yet. (Soon, though, soon…)</p>\n<p>Sharing is great, but this late into the competition, maybe we should extend the deadline a little to give the hardworking teams opportunities to flex on the shared solution compared to the spoonfeeding teams.</p>\n<p>As the final RNA test will be as late as September, a bit of extension to neutralize spoon-feeding isn't too much of a deal, is it?</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 3194904,
          "author_name": "moth",
          "author_url": "",
          "post_date": "2025-05-06T12:06:27.547000",
          "content": "<p>I am also concerned about the Early Sharing Prize. I haven't seen a deadline for it. Does it mean it will be possible to make public a high scoring notebook on the last days of the competition? This would absolutely break the LB and it is completely against Kaggle spirit. Kaggle usually disallows sharing high scoring solutions after the last two weeks of any competition.</p>",
          "votes": -1,
          "replies": [
            {
              "id": 3196666,
              "author_name": "greySnow",
              "author_url": "",
              "post_date": "2025-05-07T09:44:09.603000",
              "content": "<blockquote>\n  <p>Kaggle usually disallows sharing high scoring solutions after the last two weeks of any competition.  </p>\n</blockquote>\n<p>Only in the last week.  </p>",
              "votes": 1,
              "replies": []
            },
            {
              "id": 3196846,
              "author_name": "moth",
              "author_url": "",
              "post_date": "2025-05-07T13:49:03.207000",
              "content": "<p>Thanks, I thought it was the last 2</p>",
              "votes": 0,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 3197211,
      "author_name": "Luis Benitez",
      "author_url": "",
      "post_date": "2025-05-07T22:19:53.557000",
      "content": "<p>its just amazing!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 3194676,
      "author_name": "Nigmat Rahim",
      "author_url": "",
      "post_date": "2025-05-06T07:26:15.963000",
      "content": "<p>Encouraging news!!!</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3194234": "Hi everyone, it's great to see the rapid progress in RNA 3D structure prediction with so many participants and the sharing of novel approaches.\n\nAs we enter the final month, we have some exciting news to share.\n\nAs you may know, Nvidia has been helping @shujun17 and our lab train the 100M RibonanzaNet2 (RNet2) model.\n\nWe released the alpha version of RNet2 on Kaggle last month and the model was highlighted by Nvidia in their [developer blog](https://developer.nvidia.com/blog/stanford-das-lab-accelerates-rna-folding-research-with-nvidia-dgx-cloud/).\n\nNow, our Nvidia research collaborator @chrismunley as well as Kaggle grandmasters @youhanlee and KGMoN member @theoveil are joining as official co-hosts of this competition!\n\nYou may see them providing some interesting baseline models and answering questions on forums. We are very excited to have Nvidia's scientific collaboration in tackling this grand challenge in biology.\n",
    "3194834": "@rhijudas I’m concerned GMs might release code that break the Leaderboard, specially now that we are on the last stage of the competition. I am in favor of sharing but not spoon feeding. Its not fair for the ones who have worked hard on their submissions",
    "3197211": "its just amazing!",
    "3194676": "Encouraging news!!!"
  }
}