{
  "id": 204881,
  "title": "Will there be GPU acceleration when submitting even if I didn't use it when inference",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/204881",
  "author_name": "",
  "post_date": "2020-12-17T09:23:59.470076200Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Will there be GPU acceleration when submitting even if I didn't use it when inference?<br>\nwhen submit my inference result, whether the platform will run my notebook with GPU acceleration ?  I guess it depends on whether I used GPU acceleration when inference, because I turned off GPU in inference notebook to save GPU time, and then I found the result came out so slow when I submit.<br>\nAm I right?</p>",
  "messages": [
    {
      "id": "1116569",
      "postDate": "12/17/2020 09:23:59",
      "content": "<p>Will there be GPU acceleration when submitting even if I didn't use it when inference?<br>\nwhen submit my inference result, whether the platform will run my notebook with GPU acceleration ?  I guess it depends on whether I used GPU acceleration when inference, because I turned off GPU in inference notebook to save GPU time, and then I found the result came out so slow when I submit.<br>\nAm I right?</p>",
      "rawMarkdown": "Will there be GPU acceleration when submitting even if I didn't use it when inference?\nwhen submit my inference result, whether the platform will run my notebook with GPU acceleration ?  I guess it depends on whether I used GPU acceleration when inference, because I turned off GPU in inference notebook to save GPU time, and then I found the result came out so slow when I submit.\nAm I right?",
      "votes": null
    },
    {
      "id": "1116582",
      "postDate": "12/17/2020 09:44:02",
      "content": "<p><a href=\"https://www.kaggle.com/yihaomu\" target=\"_blank\">@yihaomu</a> If you did not turn on GPU for the inference notebook, then your submission will use only CPU even for running on the public and private test dataset when you click submit.</p>",
      "rawMarkdown": "yihaomu If you did not turn on GPU for the inference notebook, then your submission will use only CPU even for running on the public and private test dataset when you click submit.",
      "votes": null
    },
    {
      "id": "1116750",
      "postDate": "12/17/2020 12:47:25",
      "content": "<p>Indeed. If you need to save time, you can of course write your notebook with GPU off. Then you change it so that it only runs the full code when a real submission is running (and just outputs a dummy submission otherwise). That way, you can prepare it and submit it to the leaderboard without using (hardly) any of your GPU team, if you are running short. It does cut out some useful testing though.</p>",
      "rawMarkdown": "Indeed. If you need to save time, you can of course write your notebook with GPU off. Then you change it so that it only runs the full code when a real submission is running (and just outputs a dummy submission otherwise). That way, you can prepare it and submit it to the leaderboard without using (hardly) any of your GPU team, if you are running short. It does cut out some useful testing though.",
      "votes": null
    },
    {
      "id": "1116762",
      "postDate": "12/17/2020 12:58:26",
      "content": "<p>got it, thank you :)</p>",
      "rawMarkdown": "got it, thank you :)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1116582,
      "author_name": "yovinyahathugoda",
      "author_url": "",
      "post_date": "12/17/2020 09:44:02",
      "content": "<p><a href=\"https://www.kaggle.com/yihaomu\" target=\"_blank\">@yihaomu</a> If you did not turn on GPU for the inference notebook, then your submission will use only CPU even for running on the public and private test dataset when you click submit.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1116750,
          "author_name": "bjoernholzhauer",
          "author_url": "",
          "post_date": "12/17/2020 12:47:25",
          "content": "<p>Indeed. If you need to save time, you can of course write your notebook with GPU off. Then you change it so that it only runs the full code when a real submission is running (and just outputs a dummy submission otherwise). That way, you can prepare it and submit it to the leaderboard without using (hardly) any of your GPU team, if you are running short. It does cut out some useful testing though.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1116762,
          "author_name": "yihaomu",
          "author_url": "",
          "post_date": "12/17/2020 12:58:26",
          "content": "<p>got it, thank you :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1116569": "Will there be GPU acceleration when submitting even if I didn't use it when inference?\nwhen submit my inference result, whether the platform will run my notebook with GPU acceleration ?  I guess it depends on whether I used GPU acceleration when inference, because I turned off GPU in inference notebook to save GPU time, and then I found the result came out so slow when I submit.\nAm I right?",
    "1116582": "yihaomu If you did not turn on GPU for the inference notebook, then your submission will use only CPU even for running on the public and private test dataset when you click submit.",
    "1116750": "Indeed. If you need to save time, you can of course write your notebook with GPU off. Then you change it so that it only runs the full code when a real submission is running (and just outputs a dummy submission otherwise). That way, you can prepare it and submit it to the leaderboard without using (hardly) any of your GPU team, if you are running short. It does cut out some useful testing though.",
    "1116762": "got it, thank you :)"
  },
  "source": "meta"
}