{
  "id": 494753,
  "title": "Competition More focused on Optimization of Notebook",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/494753",
  "author_name": "",
  "post_date": "2024-04-18T09:23:29.004427500Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi Guys,<br>\nThis competition seems more about solving Notebook Exception and memory e errors rather than solving the Core Home Credit Problem. Why can't you allow normal submission of the output and later ask for the entire notebook of the solution to verify the approach? There's a lot of pre processing  that can be applied since so many columns are available but due to restricted memory usage not able to do it. I don't feel there's any kinda memory constraint left now in real life scenario (e.x. LLM) with ML/DL models. So, just not sure why can't we focus on getting better result  rather than focusing on optimization.</p>\n<p>Is it possible to submit Notebook that loads saved model and just run it on test data??</p>",
  "messages": [
    {
      "id": "2758654",
      "postDate": "04/18/2024 09:23:29",
      "content": "<p>Hi Guys,<br>\nThis competition seems more about solving Notebook Exception and memory e errors rather than solving the Core Home Credit Problem. Why can't you allow normal submission of the output and later ask for the entire notebook of the solution to verify the approach? There's a lot of pre processing  that can be applied since so many columns are available but due to restricted memory usage not able to do it. I don't feel there's any kinda memory constraint left now in real life scenario (e.x. LLM) with ML/DL models. So, just not sure why can't we focus on getting better result  rather than focusing on optimization.</p>\n<p>Is it possible to submit Notebook that loads saved model and just run it on test data??</p>",
      "rawMarkdown": "Hi Guys,\nThis competition seems more about solving Notebook Exception and memory e errors rather than solving the Core Home Credit Problem. Why can't you allow normal submission of the output and later ask for the entire notebook of the solution to verify the approach? There's a lot of pre processing  that can be applied since so many columns are available but due to restricted memory usage not able to do it. I don't feel there's any kinda memory constraint left now in real life scenario (e.x. LLM) with ML/DL models. So, just not sure why can't we focus on getting better result  rather than focusing on optimization.\n\n\nIs it possible to submit Notebook that loads saved model and just run it on test data??",
      "votes": null
    },
    {
      "id": "2758721",
      "postDate": "04/18/2024 10:06:33",
      "content": "<p>Yes, you can submit Notebook that loads saved model and just run it on test data</p>",
      "rawMarkdown": "Yes, you can submit Notebook that loads saved model and just run it on test data",
      "votes": null
    },
    {
      "id": "2758751",
      "postDate": "04/18/2024 10:22:51",
      "content": "<p>oh okay. Thanks, will try it.</p>",
      "rawMarkdown": "oh okay. Thanks, will try it.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2758721,
      "author_name": "fadylabib",
      "author_url": "",
      "post_date": "04/18/2024 10:06:33",
      "content": "<p>Yes, you can submit Notebook that loads saved model and just run it on test data</p>",
      "votes": null,
      "replies": [
        {
          "id": 2758751,
          "author_name": "cs17m041",
          "author_url": "",
          "post_date": "04/18/2024 10:22:51",
          "content": "<p>oh okay. Thanks, will try it.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2758654": "Hi Guys,\nThis competition seems more about solving Notebook Exception and memory e errors rather than solving the Core Home Credit Problem. Why can't you allow normal submission of the output and later ask for the entire notebook of the solution to verify the approach? There's a lot of pre processing  that can be applied since so many columns are available but due to restricted memory usage not able to do it. I don't feel there's any kinda memory constraint left now in real life scenario (e.x. LLM) with ML/DL models. So, just not sure why can't we focus on getting better result  rather than focusing on optimization.\n\n\nIs it possible to submit Notebook that loads saved model and just run it on test data??",
    "2758721": "Yes, you can submit Notebook that loads saved model and just run it on test data",
    "2758751": "oh okay. Thanks, will try it."
  },
  "source": "meta"
}