{
  "id": 290320,
  "title": "It would be much more fun if we have performance leaderboard",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/290320",
  "author_name": "",
  "post_date": "2021-11-24T04:10:12.726073700Z",
  "votes": 18,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Dear host and kaggle team,<br>\nThanks for hosting such a interesting competition! I'm very excited about performance prize in this competition.</p>\n<p>But with current setting it's almost blind about how other teams are doing well in terms of speed performance.<br>\nSo I'd like to suggest to have a leaderboard for performance prize separately so that everyone can know how good our own model's performance is.</p>\n<p>I guess it may be difficult as it requires a lot of changes, but I'm pretty sure it would definitely encourage people to develop more useful models in edge device.</p>\n<p>Thanks, I'd like to here what other people think too!</p>",
  "messages": [
    {
      "id": "1593559",
      "postDate": "11/24/2021 04:10:12",
      "content": "<p>Dear host and kaggle team,<br>\nThanks for hosting such a interesting competition! I'm very excited about performance prize in this competition.</p>\n<p>But with current setting it's almost blind about how other teams are doing well in terms of speed performance.<br>\nSo I'd like to suggest to have a leaderboard for performance prize separately so that everyone can know how good our own model's performance is.</p>\n<p>I guess it may be difficult as it requires a lot of changes, but I'm pretty sure it would definitely encourage people to develop more useful models in edge device.</p>\n<p>Thanks, I'd like to here what other people think too!</p>",
      "rawMarkdown": "Dear host and kaggle team,\nThanks for hosting such a interesting competition! I'm very excited about performance prize in this competition.\n\nBut with current setting it's almost blind about how other teams are doing well in terms of speed performance.\nSo I'd like to suggest to have a leaderboard for performance prize separately so that everyone can know how good our own model's performance is.\n\nI guess it may be difficult as it requires a lot of changes, but I'm pretty sure it would definitely encourage people to develop more useful models in edge device.\n\nThanks, I'd like to here what other people think too!",
      "votes": null
    },
    {
      "id": "1594111",
      "postDate": "11/24/2021 14:44:25",
      "content": "<p>Thanks for the feedback <a href=\"https://www.kaggle.com/bamps53\" target=\"_blank\">@bamps53</a>! I agree this would be great, but we unfortunately aren't able to make this happen here. You're right that it would take a lot of changes (and we typically do not expose runtimes in code competitions to discourage timing-based probing).</p>",
      "rawMarkdown": "Thanks for the feedback @bamps53! I agree this would be great, but we unfortunately aren't able to make this happen here. You're right that it would take a lot of changes (and we typically do not expose runtimes in code competitions to discourage timing-based probing).",
      "votes": null
    },
    {
      "id": "1594137",
      "postDate": "11/24/2021 15:04:39",
      "content": "<p><a href=\"https://www.kaggle.com/wcukierski\" target=\"_blank\">@wcukierski</a> Thanks for the answer! I understood.<br>\nLet me make sure one more thing, our models' performance will be evaluated by runtimes (and also enough accuracy to pass 10% criteria), is it correct?</p>",
      "rawMarkdown": "wcukierski Thanks for the answer! I understood.\nLet me make sure one more thing, our models' performance will be evaluated by runtimes (and also enough accuracy to pass 10% criteria), is it correct?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1594111,
      "author_name": "wcukierski",
      "author_url": "",
      "post_date": "11/24/2021 14:44:25",
      "content": "<p>Thanks for the feedback <a href=\"https://www.kaggle.com/bamps53\" target=\"_blank\">@bamps53</a>! I agree this would be great, but we unfortunately aren't able to make this happen here. You're right that it would take a lot of changes (and we typically do not expose runtimes in code competitions to discourage timing-based probing).</p>",
      "votes": null,
      "replies": [
        {
          "id": 1594137,
          "author_name": "bamps53",
          "author_url": "",
          "post_date": "11/24/2021 15:04:39",
          "content": "<p><a href=\"https://www.kaggle.com/wcukierski\" target=\"_blank\">@wcukierski</a> Thanks for the answer! I understood.<br>\nLet me make sure one more thing, our models' performance will be evaluated by runtimes (and also enough accuracy to pass 10% criteria), is it correct?</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1593559": "Dear host and kaggle team,\nThanks for hosting such a interesting competition! I'm very excited about performance prize in this competition.\n\nBut with current setting it's almost blind about how other teams are doing well in terms of speed performance.\nSo I'd like to suggest to have a leaderboard for performance prize separately so that everyone can know how good our own model's performance is.\n\nI guess it may be difficult as it requires a lot of changes, but I'm pretty sure it would definitely encourage people to develop more useful models in edge device.\n\nThanks, I'd like to here what other people think too!",
    "1594111": "Thanks for the feedback @bamps53! I agree this would be great, but we unfortunately aren't able to make this happen here. You're right that it would take a lot of changes (and we typically do not expose runtimes in code competitions to discourage timing-based probing).",
    "1594137": "wcukierski Thanks for the answer! I understood.\nLet me make sure one more thing, our models' performance will be evaluated by runtimes (and also enough accuracy to pass 10% criteria), is it correct?"
  },
  "source": "meta"
}