{
  "id": 218177,
  "title": "Is the submission time of kaggle dynamic? ",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/218177",
  "author_name": "",
  "post_date": "2021-02-09T15:07:17.545975100Z",
  "votes": 5,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Will the submission time of kaggle change with the load of the system ？ the system load is high, the submission time is long, otherwise, the submission time is short ？ I added a model to a 7-hour submission, and now it only takes 6 hours,  which is amazing</p>",
  "messages": [
    {
      "id": "1193298",
      "postDate": "02/09/2021 15:07:17",
      "content": "<p>Will the submission time of kaggle change with the load of the system ？ the system load is high, the submission time is long, otherwise, the submission time is short ？ I added a model to a 7-hour submission, and now it only takes 6 hours,  which is amazing</p>",
      "rawMarkdown": "Will the submission time of kaggle change with the load of the system ？ the system load is high, the submission time is long, otherwise, the submission time is short ？ I added a model to a 7-hour submission, and now it only takes 6 hours,  which is amazing",
      "votes": null
    },
    {
      "id": "1193371",
      "postDate": "02/09/2021 15:52:28",
      "content": "<p>Yes. I have noticed this too. <br>\nFirstly, the Kaggle clock rounds time in a simple way. if your model is running sine 6 hours and 31 minutes it rounds it off to 7 while anything less than 6:30 would be shown as 6.<br>\nAlso, the time is indeed dynamic. Sometimes my kernels that are Submitted later are executed faster than the ones submitted earlier the only difference in them being the ensembling parameters.</p>",
      "rawMarkdown": "Yes. I have noticed this too. \nFirstly, the Kaggle clock rounds time in a simple way. if your model is running sine 6 hours and 31 minutes it rounds it off to 7 while anything less than 6:30 would be shown as 6.\nAlso, the time is indeed dynamic. Sometimes my kernels that are Submitted later are executed faster than the ones submitted earlier the only difference in them being the ensembling parameters.",
      "votes": null
    },
    {
      "id": "1193407",
      "postDate": "02/09/2021 16:14:48",
      "content": "<p>Same.<br>\nLost many subs because of this.</p>",
      "rawMarkdown": "Same.\nLost many subs because of this.",
      "votes": null
    },
    {
      "id": "1205634",
      "postDate": "02/16/2021 20:41:16",
      "content": "<p>This is getting ridiculous now. Can't add a model because it 'might' exceed time limit and waste a submission. And idk if it's about the load on the system, I made 2 submissions together, one is 25-30 minutes longer than the other, but it has been 90 minutes and the kernel hasn't stopped executing. </p>",
      "rawMarkdown": "This is getting ridiculous now. Can't add a model because it 'might' exceed time limit and waste a submission. And idk if it's about the load on the system, I made 2 submissions together, one is 25-30 minutes longer than the other, but it has been 90 minutes and the kernel hasn't stopped executing.",
      "votes": null
    },
    {
      "id": "1206008",
      "postDate": "02/17/2021 05:31:56",
      "content": "<p>I'm not sure. I'm just guessing</p>",
      "rawMarkdown": "I'm not sure. I'm just guessing",
      "votes": null
    },
    {
      "id": "1206503",
      "postDate": "02/17/2021 11:53:18",
      "content": "<p>I have noticed issues like this in past competitions too. I have an impression that the exact same submission can take longer when Kaggle servers are overloaded, which is likely to happen in the last 1-2 days of the competition. Therefore, I guess we need to be a little conservative with the number of models in the last days…</p>",
      "rawMarkdown": "I have noticed issues like this in past competitions too. I have an impression that the exact same submission can take longer when Kaggle servers are overloaded, which is likely to happen in the last 1-2 days of the competition. Therefore, I guess we need to be a little conservative with the number of models in the last days...",
      "votes": null
    },
    {
      "id": "1206517",
      "postDate": "02/17/2021 12:01:40",
      "content": "<p>We need to be as careful as possible to allow an hour of redundancy</p>",
      "rawMarkdown": "We need to be as careful as possible to allow an hour of redundancy",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1193371,
      "author_name": "namantuli",
      "author_url": "",
      "post_date": "02/09/2021 15:52:28",
      "content": "<p>Yes. I have noticed this too. <br>\nFirstly, the Kaggle clock rounds time in a simple way. if your model is running sine 6 hours and 31 minutes it rounds it off to 7 while anything less than 6:30 would be shown as 6.<br>\nAlso, the time is indeed dynamic. Sometimes my kernels that are Submitted later are executed faster than the ones submitted earlier the only difference in them being the ensembling parameters.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1193407,
      "author_name": "woshifym",
      "author_url": "",
      "post_date": "02/09/2021 16:14:48",
      "content": "<p>Same.<br>\nLost many subs because of this.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1205634,
      "author_name": "albernard",
      "author_url": "",
      "post_date": "02/16/2021 20:41:16",
      "content": "<p>This is getting ridiculous now. Can't add a model because it 'might' exceed time limit and waste a submission. And idk if it's about the load on the system, I made 2 submissions together, one is 25-30 minutes longer than the other, but it has been 90 minutes and the kernel hasn't stopped executing. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1206008,
          "author_name": "hanson0910",
          "author_url": "",
          "post_date": "02/17/2021 05:31:56",
          "content": "<p>I'm not sure. I'm just guessing</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1206503,
      "author_name": "kozodoi",
      "author_url": "",
      "post_date": "02/17/2021 11:53:18",
      "content": "<p>I have noticed issues like this in past competitions too. I have an impression that the exact same submission can take longer when Kaggle servers are overloaded, which is likely to happen in the last 1-2 days of the competition. Therefore, I guess we need to be a little conservative with the number of models in the last days…</p>",
      "votes": null,
      "replies": [
        {
          "id": 1206517,
          "author_name": "hanson0910",
          "author_url": "",
          "post_date": "02/17/2021 12:01:40",
          "content": "<p>We need to be as careful as possible to allow an hour of redundancy</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1193298": "Will the submission time of kaggle change with the load of the system ？ the system load is high, the submission time is long, otherwise, the submission time is short ？ I added a model to a 7-hour submission, and now it only takes 6 hours,  which is amazing",
    "1193371": "Yes. I have noticed this too. \nFirstly, the Kaggle clock rounds time in a simple way. if your model is running sine 6 hours and 31 minutes it rounds it off to 7 while anything less than 6:30 would be shown as 6.\nAlso, the time is indeed dynamic. Sometimes my kernels that are Submitted later are executed faster than the ones submitted earlier the only difference in them being the ensembling parameters.",
    "1193407": "Same.\nLost many subs because of this.",
    "1205634": "This is getting ridiculous now. Can't add a model because it 'might' exceed time limit and waste a submission. And idk if it's about the load on the system, I made 2 submissions together, one is 25-30 minutes longer than the other, but it has been 90 minutes and the kernel hasn't stopped executing.",
    "1206008": "I'm not sure. I'm just guessing",
    "1206503": "I have noticed issues like this in past competitions too. I have an impression that the exact same submission can take longer when Kaggle servers are overloaded, which is likely to happen in the last 1-2 days of the competition. Therefore, I guess we need to be a little conservative with the number of models in the last days...",
    "1206517": "We need to be as careful as possible to allow an hour of redundancy"
  },
  "source": "meta"
}