{
  "id": 74340,
  "title": "Confusion regarding execution time limit",
  "url": "/competitions/quora-insincere-questions-classification/discussion/74340",
  "author_name": "",
  "post_date": "2018-12-11T11:42:00.479205900Z",
  "votes": 8,
  "comment_count": 7,
  "views": 0,
  "content": "<p>Not sure if I am doing something wrong, but after I forked public kernel and made minor changes, it achieved score of 0.701 on public LB which surprised me. Then I checked and noticed that kernel was running for 8220.5s on GPU which exceeds execution time limit. In Kernels FAQ it says:</p>\n\n<p>&gt; GPUs are enabled for this competition. If you use GPUs, you will be limited to 2 hours of run time. If you do not use GPUs, you will be limited to 6 hours of run time. If you attempt to make a submission whose kernel exceeds these limits, you will receive an error.</p>\n\n<p>Whereas I didn't get any error.\nIs it me or is this something to be concerned about? </p>",
  "messages": [
    {
      "id": "437110",
      "postDate": "12/11/2018 11:42:00",
      "content": "<p>Not sure if I am doing something wrong, but after I forked public kernel and made minor changes, it achieved score of 0.701 on public LB which surprised me. Then I checked and noticed that kernel was running for 8220.5s on GPU which exceeds execution time limit. In Kernels FAQ it says:</p>\n\n<p>&gt; GPUs are enabled for this competition. If you use GPUs, you will be limited to 2 hours of run time. If you do not use GPUs, you will be limited to 6 hours of run time. If you attempt to make a submission whose kernel exceeds these limits, you will receive an error.</p>\n\n<p>Whereas I didn't get any error.\nIs it me or is this something to be concerned about? </p>",
      "rawMarkdown": "Not sure if I am doing something wrong, but after I forked public kernel and made minor changes, it achieved score of 0.701 on public LB which surprised me. Then I checked and noticed that kernel was running for 8220.5s on GPU which exceeds execution time limit. In Kernels FAQ it says:\n\n&gt; GPUs are enabled for this competition. If you use GPUs, you will be limited to 2 hours of run time. If you do not use GPUs, you will be limited to 6 hours of run time. If you attempt to make a submission whose kernel exceeds these limits, you will receive an error.\n\nWhereas I didn't get any error.\nIs it me or is this something to be concerned about?",
      "votes": null
    },
    {
      "id": "437271",
      "postDate": "12/11/2018 16:47:02",
      "content": "<p>Probably a bug.. </p>\n\n<p>Anyway , make sure your kernels run in less than 2hrs .., given kaggle will run again every selected kernel on new test set during the 2nd stage.   </p>\n\n<p>Many teams didn't get ranked during Mercari competition , given their kernels didn't run successfully within the time constraint and RAM limit during the 2nd stage</p>",
      "rawMarkdown": "Probably a bug.. \n\nAnyway , make sure your kernels run in less than 2hrs .., given kaggle will run again every selected kernel on new test set during the 2nd stage.   \n\nMany teams didn't get ranked during Mercari competition , given their kernels didn't run successfully within the time constraint and RAM limit during the 2nd stage",
      "votes": null
    },
    {
      "id": "437281",
      "postDate": "12/11/2018 17:08:58",
      "content": "<p>Thank you! :)\nI just thought it could be misleading since it's definitely not valid submission then.</p>",
      "rawMarkdown": "Thank you! :)\nI just thought it could be misleading since it's definitely not valid submission then.",
      "votes": null
    },
    {
      "id": "437358",
      "postDate": "12/11/2018 19:09:29",
      "content": "<p>I tried it too. My submission with 8200+ seconds got submitted. But, alas that didn't score 0.701 ;) </p>",
      "rawMarkdown": "I tried it too. My submission with 8200+ seconds got submitted. But, alas that didn't score 0.701 ;)",
      "votes": null
    },
    {
      "id": "437415",
      "postDate": "12/11/2018 21:36:54",
      "content": "<p>I had execution times a few seconds too long and I could not submit, so I really would make sure it runs within the time limit.</p>",
      "rawMarkdown": "I had execution times a few seconds too long and I could not submit, so I really would make sure it runs within the time limit.",
      "votes": null
    },
    {
      "id": "437577",
      "postDate": "12/12/2018 06:33:24",
      "content": "<p>I also find this problem, from yesterday, this competition may lost time limit.</p>",
      "rawMarkdown": "I also find this problem, from yesterday, this competition may lost time limit.",
      "votes": null
    },
    {
      "id": "438050",
      "postDate": "12/13/2018 02:48:34",
      "content": "<p>minor changes of public kernel, amazing result! Is it reproducible?</p>",
      "rawMarkdown": "minor changes of public kernel, amazing result! Is it reproducible?",
      "votes": null
    },
    {
      "id": "438207",
      "postDate": "12/13/2018 09:12:13",
      "content": "<p>Nop, it's in Keras and I didn't do anything about it, but I ran it two more times as sanity check and the other two submissions got 0.700 and 0.698 </p>",
      "rawMarkdown": "Nop, it's in Keras and I didn't do anything about it, but I ran it two more times as sanity check and the other two submissions got 0.700 and 0.698",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 437271,
      "author_name": "serigne",
      "author_url": "",
      "post_date": "12/11/2018 16:47:02",
      "content": "<p>Probably a bug.. </p>\n\n<p>Anyway , make sure your kernels run in less than 2hrs .., given kaggle will run again every selected kernel on new test set during the 2nd stage.   </p>\n\n<p>Many teams didn't get ranked during Mercari competition , given their kernels didn't run successfully within the time constraint and RAM limit during the 2nd stage</p>",
      "votes": null,
      "replies": [
        {
          "id": 437281,
          "author_name": "mihajlot",
          "author_url": "",
          "post_date": "12/11/2018 17:08:58",
          "content": "<p>Thank you! :)\nI just thought it could be misleading since it's definitely not valid submission then.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 437358,
      "author_name": "shaz13",
      "author_url": "",
      "post_date": "12/11/2018 19:09:29",
      "content": "<p>I tried it too. My submission with 8200+ seconds got submitted. But, alas that didn't score 0.701 ;) </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 437415,
      "author_name": "philippsinger",
      "author_url": "",
      "post_date": "12/11/2018 21:36:54",
      "content": "<p>I had execution times a few seconds too long and I could not submit, so I really would make sure it runs within the time limit.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 437577,
      "author_name": "guang10",
      "author_url": "",
      "post_date": "12/12/2018 06:33:24",
      "content": "<p>I also find this problem, from yesterday, this competition may lost time limit.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 438050,
      "author_name": "zjucor",
      "author_url": "",
      "post_date": "12/13/2018 02:48:34",
      "content": "<p>minor changes of public kernel, amazing result! Is it reproducible?</p>",
      "votes": null,
      "replies": [
        {
          "id": 438207,
          "author_name": "mihajlot",
          "author_url": "",
          "post_date": "12/13/2018 09:12:13",
          "content": "<p>Nop, it's in Keras and I didn't do anything about it, but I ran it two more times as sanity check and the other two submissions got 0.700 and 0.698 </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "437110": "Not sure if I am doing something wrong, but after I forked public kernel and made minor changes, it achieved score of 0.701 on public LB which surprised me. Then I checked and noticed that kernel was running for 8220.5s on GPU which exceeds execution time limit. In Kernels FAQ it says:\n\n&gt; GPUs are enabled for this competition. If you use GPUs, you will be limited to 2 hours of run time. If you do not use GPUs, you will be limited to 6 hours of run time. If you attempt to make a submission whose kernel exceeds these limits, you will receive an error.\n\nWhereas I didn't get any error.\nIs it me or is this something to be concerned about?",
    "437271": "Probably a bug.. \n\nAnyway , make sure your kernels run in less than 2hrs .., given kaggle will run again every selected kernel on new test set during the 2nd stage.   \n\nMany teams didn't get ranked during Mercari competition , given their kernels didn't run successfully within the time constraint and RAM limit during the 2nd stage",
    "437281": "Thank you! :)\nI just thought it could be misleading since it's definitely not valid submission then.",
    "437358": "I tried it too. My submission with 8200+ seconds got submitted. But, alas that didn't score 0.701 ;)",
    "437415": "I had execution times a few seconds too long and I could not submit, so I really would make sure it runs within the time limit.",
    "437577": "I also find this problem, from yesterday, this competition may lost time limit.",
    "438050": "minor changes of public kernel, amazing result! Is it reproducible?",
    "438207": "Nop, it's in Keras and I didn't do anything about it, but I ran it two more times as sanity check and the other two submissions got 0.700 and 0.698"
  },
  "source": "meta"
}