{
  "id": 552024,
  "title": "Scoring runs forever",
  "url": "/competitions/child-mind-institute-problematic-internet-use/discussion/552024",
  "author_name": "",
  "post_date": "2024-12-17T07:17:47.472524700Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hello everyone,<br>\nmy latest submission has been scoring for 9 hours by now. I didn't change alot, so I assume the problem is at kaggles end. Should I just resubmit the same code? I don't want to waste another submission if it gets stuck again. Is that a common problem that comes up from time to time? I saw some similar posts years ago, but no answere to them.</p>\n<p>Edit: Tagging staff member <a href=\"https://www.kaggle.com/sohier\" target=\"_blank\">@sohier</a> </p>",
  "messages": [
    {
      "id": "3073981",
      "postDate": "12/17/2024 07:17:47",
      "content": "<p>Hello everyone,<br>\nmy latest submission has been scoring for 9 hours by now. I didn't change alot, so I assume the problem is at kaggles end. Should I just resubmit the same code? I don't want to waste another submission if it gets stuck again. Is that a common problem that comes up from time to time? I saw some similar posts years ago, but no answere to them.</p>\n<p>Edit: Tagging staff member <a href=\"https://www.kaggle.com/sohier\" target=\"_blank\">@sohier</a> </p>",
      "rawMarkdown": "Hello everyone,\nmy latest submission has been scoring for 9 hours by now. I didn't change alot, so I assume the problem is at kaggles end. Should I just resubmit the same code? I don't want to waste another submission if it gets stuck again. Is that a common problem that comes up from time to time? I saw some similar posts years ago, but no answere to them.\n\nEdit: Tagging staff member @sohier",
      "votes": null
    },
    {
      "id": "3073988",
      "postDate": "12/17/2024 07:30:25",
      "content": "<p><a href=\"https://www.kaggle.com/mariusheuser\" target=\"_blank\">@mariusheuser</a> you can raise this to Kaggle and perhaps tag the relevant Kaggle staff member in charge of this competition. I think <a href=\"https://www.kaggle.com/sohier\" target=\"_blank\">@sohier</a> is your contact point here. </p>",
      "rawMarkdown": "mariusheuser you can raise this to Kaggle and perhaps tag the relevant Kaggle staff member in charge of this competition. I think @sohier is your contact point here.",
      "votes": null
    },
    {
      "id": "3074902",
      "postDate": "12/18/2024 07:37:57",
      "content": "<p>This problem happend yesterday again. There seems to be something wrong with my code now, sometimes. Because some went through. It's really annoying to lose predictions in the final stage of the competition after investing so much time. Is there any way to directly contact kaggle if the problem is on my end or theirs?<br>\nThe only big thing I changed recently is my tabnet model, which I saved and now just manually load in every submit.</p>",
      "rawMarkdown": "This problem happend yesterday again. There seems to be something wrong with my code now, sometimes. Because some went through. It's really annoying to lose predictions in the final stage of the competition after investing so much time. Is there any way to directly contact kaggle if the problem is on my end or theirs?\nThe only big thing I changed recently is my tabnet model, which I saved and now just manually load in every submit.",
      "votes": null
    },
    {
      "id": "3075147",
      "postDate": "12/18/2024 15:20:55",
      "content": "<p>The issue seems to be TabNet. If I use it for my prediction, the scoring runs forever.<br>\nI used tabnet before and didn't change alot. I changed model parameters. Seems like something is broken there.<br>\nAnyone ever experienced such a problem?<br>\nI save the .json and network and upload it as dataset and then move it back to kaggle/working/ to load it in my submission.<br>\nAny help would be appreciated 🙏</p>",
      "rawMarkdown": "The issue seems to be TabNet. If I use it for my prediction, the scoring runs forever.\nI used tabnet before and didn't change alot. I changed model parameters. Seems like something is broken there.\nAnyone ever experienced such a problem?\nI save the .json and network and upload it as dataset and then move it back to kaggle/working/ to load it in my submission.\nAny help would be appreciated 🙏",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3073988,
      "author_name": "ravi20076",
      "author_url": "",
      "post_date": "12/17/2024 07:30:25",
      "content": "<p><a href=\"https://www.kaggle.com/mariusheuser\" target=\"_blank\">@mariusheuser</a> you can raise this to Kaggle and perhaps tag the relevant Kaggle staff member in charge of this competition. I think <a href=\"https://www.kaggle.com/sohier\" target=\"_blank\">@sohier</a> is your contact point here. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3074902,
      "author_name": "mariusheuser",
      "author_url": "",
      "post_date": "12/18/2024 07:37:57",
      "content": "<p>This problem happend yesterday again. There seems to be something wrong with my code now, sometimes. Because some went through. It's really annoying to lose predictions in the final stage of the competition after investing so much time. Is there any way to directly contact kaggle if the problem is on my end or theirs?<br>\nThe only big thing I changed recently is my tabnet model, which I saved and now just manually load in every submit.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3075147,
      "author_name": "mariusheuser",
      "author_url": "",
      "post_date": "12/18/2024 15:20:55",
      "content": "<p>The issue seems to be TabNet. If I use it for my prediction, the scoring runs forever.<br>\nI used tabnet before and didn't change alot. I changed model parameters. Seems like something is broken there.<br>\nAnyone ever experienced such a problem?<br>\nI save the .json and network and upload it as dataset and then move it back to kaggle/working/ to load it in my submission.<br>\nAny help would be appreciated 🙏</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3073981": "Hello everyone,\nmy latest submission has been scoring for 9 hours by now. I didn't change alot, so I assume the problem is at kaggles end. Should I just resubmit the same code? I don't want to waste another submission if it gets stuck again. Is that a common problem that comes up from time to time? I saw some similar posts years ago, but no answere to them.\n\nEdit: Tagging staff member @sohier",
    "3073988": "mariusheuser you can raise this to Kaggle and perhaps tag the relevant Kaggle staff member in charge of this competition. I think @sohier is your contact point here.",
    "3074902": "This problem happend yesterday again. There seems to be something wrong with my code now, sometimes. Because some went through. It's really annoying to lose predictions in the final stage of the competition after investing so much time. Is there any way to directly contact kaggle if the problem is on my end or theirs?\nThe only big thing I changed recently is my tabnet model, which I saved and now just manually load in every submit.",
    "3075147": "The issue seems to be TabNet. If I use it for my prediction, the scoring runs forever.\nI used tabnet before and didn't change alot. I changed model parameters. Seems like something is broken there.\nAnyone ever experienced such a problem?\nI save the .json and network and upload it as dataset and then move it back to kaggle/working/ to load it in my submission.\nAny help would be appreciated 🙏"
  },
  "source": "meta"
}