{
  "id": 681267,
  "title": "Scoring related Issue. ",
  "url": "/competitions/stanford-rna-3d-folding-2/discussion/681267",
  "author_name": "Advait ithape",
  "post_date": "2026-03-13T18:21:48.852000",
  "votes": 0,
  "comment_count": 6,
  "views": 0,
  "content": "<p>hello everyone. i am facing an issue with the scoring results. one of my submissions has a lb score of <strong>0.426</strong>. i tried to resubmit the same notebook twice, the first time it got scored <strong>0.395</strong>, while the second time it got scored <strong>0.400</strong>. And the third time now it is been scored at <strong>0.411</strong>.</p>\n<p>Is anyone else facing similar issue. I would like to hear from the hosts, on this issue.</p>",
  "messages": [
    {
      "id": 3423700,
      "postDate": "2026-03-18T22:38:55.423Z",
      "content": "<p>I'm not sure it is even possible in most cases to get repeatable result on GPU.</p>",
      "rawMarkdown": "I'm not sure it is even possible in most cases to get repeatable result on GPU.",
      "replies": [
        {
          "id": 3423830,
          "postDate": "2026-03-19T03:59:00.310Z",
          "content": "<p>yes you can</p>",
          "rawMarkdown": "yes you can"
        }
      ]
    },
    {
      "id": 3420679,
      "postDate": "2026-03-13T18:31:53.587Z",
      "content": "<p>set seed properly</p>",
      "rawMarkdown": "set seed properly",
      "replies": [
        {
          "id": 3420682,
          "postDate": "2026-03-13T18:34:02.233Z",
          "content": "<p>seed is 42 consistent through all submissions</p>",
          "rawMarkdown": "seed is 42 consistent through all submissions",
          "replies": [
            {
              "id": 3420694,
              "postDate": "2026-03-13T18:53:55.727Z",
              "content": "<p>you set seed=42 but you also need to make sure every place where randomness exists that seed=42 is actually working… setting a seed at the top of your notebook doesn't mean you are done</p>",
              "rawMarkdown": "you set seed=42 but you also need to make sure every place where randomness exists that seed=42 is actually working... setting a seed at the top of your notebook doesn't mean you are done",
              "votes": 1
            },
            {
              "id": 3421625,
              "postDate": "2026-03-16T00:38:22.320Z",
              "content": "<p>I have tried fixing the seed contamination issue, but it consistently is giving different results. Can you suggest me exactly what to do. </p>",
              "rawMarkdown": "I have tried fixing the seed contamination issue, but it consistently is giving different results. Can you suggest me exactly what to do. "
            }
          ]
        }
      ]
    },
    {
      "id": 3420676,
      "postDate": "2026-03-13T18:21:48.853Z",
      "content": "<p>hello everyone. i am facing an issue with the scoring results. one of my submissions has a lb score of <strong>0.426</strong>. i tried to resubmit the same notebook twice, the first time it got scored <strong>0.395</strong>, while the second time it got scored <strong>0.400</strong>. And the third time now it is been scored at <strong>0.411</strong>.</p>\n<p>Is anyone else facing similar issue. I would like to hear from the hosts, on this issue.</p>",
      "rawMarkdown": "hello everyone. i am facing an issue with the scoring results. one of my submissions has a lb score of **0.426**. i tried to resubmit the same notebook twice, the first time it got scored **0.395**, while the second time it got scored **0.400**. And the third time now it is been scored at **0.411**.\n\nIs anyone else facing similar issue. I would like to hear from the hosts, on this issue."
    }
  ],
  "comments": [
    {
      "id": 3423700,
      "author_name": "Serhii Hrynko",
      "author_url": "",
      "post_date": "2026-03-18T22:38:55.423000",
      "content": "<p>I'm not sure it is even possible in most cases to get repeatable result on GPU.</p>",
      "votes": 0,
      "replies": [
        {
          "id": 3423830,
          "author_name": "hongan",
          "author_url": "",
          "post_date": "2026-03-19T03:59:00.310000",
          "content": "<p>yes you can</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 3420679,
      "author_name": "hongan",
      "author_url": "",
      "post_date": "2026-03-13T18:31:53.587000",
      "content": "<p>set seed properly</p>",
      "votes": 0,
      "replies": [
        {
          "id": 3420682,
          "author_name": "Advait ithape",
          "author_url": "",
          "post_date": "2026-03-13T18:34:02.233000",
          "content": "<p>seed is 42 consistent through all submissions</p>",
          "votes": 0,
          "replies": [
            {
              "id": 3420694,
              "author_name": "hongan",
              "author_url": "",
              "post_date": "2026-03-13T18:53:55.727000",
              "content": "<p>you set seed=42 but you also need to make sure every place where randomness exists that seed=42 is actually working… setting a seed at the top of your notebook doesn't mean you are done</p>",
              "votes": 1,
              "replies": []
            },
            {
              "id": 3421625,
              "author_name": "Advait ithape",
              "author_url": "",
              "post_date": "2026-03-16T00:38:22.320000",
              "content": "<p>I have tried fixing the seed contamination issue, but it consistently is giving different results. Can you suggest me exactly what to do. </p>",
              "votes": 0,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3423700": "I'm not sure it is even possible in most cases to get repeatable result on GPU.",
    "3420679": "set seed properly",
    "3420676": "hello everyone. i am facing an issue with the scoring results. one of my submissions has a lb score of **0.426**. i tried to resubmit the same notebook twice, the first time it got scored **0.395**, while the second time it got scored **0.400**. And the third time now it is been scored at **0.411**.\n\nIs anyone else facing similar issue. I would like to hear from the hosts, on this issue."
  }
}