{
  "id": 218146,
  "title": "A question about submission eligibility",
  "url": "/competitions/rfcx-species-audio-detection/discussion/218146",
  "author_name": "",
  "post_date": "2021-02-09T12:39:14.162653700Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I have found that my model architecture takes around 8 times longer to run on Kaggle than on my local machine, about 7 hours in total. This limits me to around 6 submissions per week with my current GPU quota.</p>\n<p>Is it permissible to simply upload the submission file, but within a notebook that also contains the code that was used? Obviously the actual prediction stage would be commented out, but it would demonstrate that the results are reproducible.</p>",
  "messages": [
    {
      "id": "1193052",
      "postDate": "02/09/2021 12:39:14",
      "content": "<p>I have found that my model architecture takes around 8 times longer to run on Kaggle than on my local machine, about 7 hours in total. This limits me to around 6 submissions per week with my current GPU quota.</p>\n<p>Is it permissible to simply upload the submission file, but within a notebook that also contains the code that was used? Obviously the actual prediction stage would be commented out, but it would demonstrate that the results are reproducible.</p>",
      "rawMarkdown": "I have found that my model architecture takes around 8 times longer to run on Kaggle than on my local machine, about 7 hours in total. This limits me to around 6 submissions per week with my current GPU quota.\n\nIs it permissible to simply upload the submission file, but within a notebook that also contains the code that was used? Obviously the actual prediction stage would be commented out, but it would demonstrate that the results are reproducible.",
      "votes": null
    },
    {
      "id": "1193066",
      "postDate": "02/09/2021 12:43:59",
      "content": "<p>You don't even have to bother with the code, uploading the submission file is enough for this competition.</p>\n<p>Regarding reproducibility, you don't have to bother with that unless you are in the prize zone :)</p>",
      "rawMarkdown": "You don't even have to bother with the code, uploading the submission file is enough for this competition.\n\nRegarding reproducibility, you don't have to bother with that unless you are in the prize zone :)",
      "votes": null
    },
    {
      "id": "1193083",
      "postDate": "02/09/2021 12:59:05",
      "content": "<p>Thanks for the clarification! Well, I probably don't have to worry about that…</p>",
      "rawMarkdown": "Thanks for the clarification! Well, I probably don't have to worry about that...",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1193066,
      "author_name": "theoviel",
      "author_url": "",
      "post_date": "02/09/2021 12:43:59",
      "content": "<p>You don't even have to bother with the code, uploading the submission file is enough for this competition.</p>\n<p>Regarding reproducibility, you don't have to bother with that unless you are in the prize zone :)</p>",
      "votes": null,
      "replies": [
        {
          "id": 1193083,
          "author_name": "bigironsphere",
          "author_url": "",
          "post_date": "02/09/2021 12:59:05",
          "content": "<p>Thanks for the clarification! Well, I probably don't have to worry about that…</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1193052": "I have found that my model architecture takes around 8 times longer to run on Kaggle than on my local machine, about 7 hours in total. This limits me to around 6 submissions per week with my current GPU quota.\n\nIs it permissible to simply upload the submission file, but within a notebook that also contains the code that was used? Obviously the actual prediction stage would be commented out, but it would demonstrate that the results are reproducible.",
    "1193066": "You don't even have to bother with the code, uploading the submission file is enough for this competition.\n\nRegarding reproducibility, you don't have to bother with that unless you are in the prize zone :)",
    "1193083": "Thanks for the clarification! Well, I probably don't have to worry about that..."
  },
  "source": "meta"
}