{
  "id": 366379,
  "title": "Have you selected the best submission?",
  "url": "/competitions/open-problems-multimodal/discussion/366379",
  "author_name": "",
  "post_date": "2022-11-16T00:47:52.581843200Z",
  "votes": 2,
  "comment_count": 8,
  "views": 0,
  "content": "<p>Our team's best private score is 0.772 but we didn't select it haha. How about you?</p>",
  "messages": [
    {
      "id": "2031247",
      "postDate": "11/16/2022 00:47:52",
      "content": "<p>Our team's best private score is 0.772 but we didn't select it haha. How about you?</p>",
      "rawMarkdown": "Our team's best private score is 0.772 but we didn't select it haha. How about you?",
      "votes": null
    },
    {
      "id": "2031248",
      "postDate": "11/16/2022 00:49:18",
      "content": "<p>Yep, we've selected two best private subs.</p>\n<p>The first one (0.771) is our best public score (0.814), the second one (0.772) had 0.808 on public lb. So trust your cv as always)</p>",
      "rawMarkdown": "Yep, we've selected two best private subs.\n\nThe first one (0.771) is our best public score (0.814), the second one (0.772) had 0.808 on public lb. So trust your cv as always)",
      "votes": null
    },
    {
      "id": "2031250",
      "postDate": "11/16/2022 00:52:22",
      "content": "<p>I selected the best LB model(on donor) and the same model as the best LB(on donor and day). They don't have much difference in private LB.</p>",
      "rawMarkdown": "I selected the best LB model(on donor) and the same model as the best LB(on donor and day). They don't have much difference in private LB.",
      "votes": null
    },
    {
      "id": "2031251",
      "postDate": "11/16/2022 00:53:14",
      "content": "<p>Wow 0.808-&gt;0.772, congratulations!</p>",
      "rawMarkdown": "Wow 0.808->0.772, congratulations!",
      "votes": null
    },
    {
      "id": "2031258",
      "postDate": "11/16/2022 00:59:32",
      "content": "<p>no,we missed it,I selected the third one :0.770😅</p>",
      "rawMarkdown": "no,we missed it,I selected the third one :0.770😅",
      "votes": null
    },
    {
      "id": "2031260",
      "postDate": "11/16/2022 01:01:47",
      "content": "<p>Congrats! Our team didn't choose the best private(also the best cv) because that is 0.813 and the one we choose is 0.814(it is hard to measure how far between these two on the public), but it is still not too shake. I guess we still need trust cv haha</p>",
      "rawMarkdown": "Congrats! Our team didn't choose the best private(also the best cv) because that is 0.813 and the one we choose is 0.814(it is hard to measure how far between these two on the public), but it is still not too shake. I guess we still need trust cv haha",
      "votes": null
    },
    {
      "id": "2031285",
      "postDate": "11/16/2022 01:37:22",
      "content": "<p>everytime should trust your cv😂</p>",
      "rawMarkdown": "everytime should trust your cv😂",
      "votes": null
    },
    {
      "id": "2031289",
      "postDate": "11/16/2022 01:47:24",
      "content": "<p>I think the shake is not super huge compared to some other comps. we just missed 1 top sub, so it is good. Also, I am guessing the random sampling evaluation of the multi-part contributed some of the shakes as well?</p>",
      "rawMarkdown": "I think the shake is not super huge compared to some other comps. we just missed 1 top sub, so it is good. Also, I am guessing the random sampling evaluation of the multi-part contributed some of the shakes as well?",
      "votes": null
    },
    {
      "id": "2031293",
      "postDate": "11/16/2022 01:55:10",
      "content": "<p>My best private score is 0.771 which was used your <a href=\"https://www.kaggle.com/code/takanashihumbert/binarized-data-for-multiome-part\" target=\"_blank\">binarized data notebook</a> (Great work, thank you so much), but it's a little worse than my selected submission on both CV and public LB, so I didn't select it.</p>",
      "rawMarkdown": "My best private score is 0.771 which was used your [binarized data notebook](https://www.kaggle.com/code/takanashihumbert/binarized-data-for-multiome-part) (Great work, thank you so much), but it's a little worse than my selected submission on both CV and public LB, so I didn't select it.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2031248,
      "author_name": "l0glikelihood",
      "author_url": "",
      "post_date": "11/16/2022 00:49:18",
      "content": "<p>Yep, we've selected two best private subs.</p>\n<p>The first one (0.771) is our best public score (0.814), the second one (0.772) had 0.808 on public lb. So trust your cv as always)</p>",
      "votes": null,
      "replies": [
        {
          "id": 2031251,
          "author_name": "takanashihumbert",
          "author_url": "",
          "post_date": "11/16/2022 00:53:14",
          "content": "<p>Wow 0.808-&gt;0.772, congratulations!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 2031285,
          "author_name": "huangzchao",
          "author_url": "",
          "post_date": "11/16/2022 01:37:22",
          "content": "<p>everytime should trust your cv😂</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2031250,
      "author_name": "qqzzxxdd",
      "author_url": "",
      "post_date": "11/16/2022 00:52:22",
      "content": "<p>I selected the best LB model(on donor) and the same model as the best LB(on donor and day). They don't have much difference in private LB.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2031258,
      "author_name": "kunmingxie",
      "author_url": "",
      "post_date": "11/16/2022 00:59:32",
      "content": "<p>no,we missed it,I selected the third one :0.770😅</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2031260,
      "author_name": "learnmore1",
      "author_url": "",
      "post_date": "11/16/2022 01:01:47",
      "content": "<p>Congrats! Our team didn't choose the best private(also the best cv) because that is 0.813 and the one we choose is 0.814(it is hard to measure how far between these two on the public), but it is still not too shake. I guess we still need trust cv haha</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2031289,
      "author_name": "kingychiu",
      "author_url": "",
      "post_date": "11/16/2022 01:47:24",
      "content": "<p>I think the shake is not super huge compared to some other comps. we just missed 1 top sub, so it is good. Also, I am guessing the random sampling evaluation of the multi-part contributed some of the shakes as well?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2031293,
      "author_name": "ingffwe",
      "author_url": "",
      "post_date": "11/16/2022 01:55:10",
      "content": "<p>My best private score is 0.771 which was used your <a href=\"https://www.kaggle.com/code/takanashihumbert/binarized-data-for-multiome-part\" target=\"_blank\">binarized data notebook</a> (Great work, thank you so much), but it's a little worse than my selected submission on both CV and public LB, so I didn't select it.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2031247": "Our team's best private score is 0.772 but we didn't select it haha. How about you?",
    "2031248": "Yep, we've selected two best private subs.\n\nThe first one (0.771) is our best public score (0.814), the second one (0.772) had 0.808 on public lb. So trust your cv as always)",
    "2031250": "I selected the best LB model(on donor) and the same model as the best LB(on donor and day). They don't have much difference in private LB.",
    "2031251": "Wow 0.808->0.772, congratulations!",
    "2031258": "no,we missed it,I selected the third one :0.770😅",
    "2031260": "Congrats! Our team didn't choose the best private(also the best cv) because that is 0.813 and the one we choose is 0.814(it is hard to measure how far between these two on the public), but it is still not too shake. I guess we still need trust cv haha",
    "2031285": "everytime should trust your cv😂",
    "2031289": "I think the shake is not super huge compared to some other comps. we just missed 1 top sub, so it is good. Also, I am guessing the random sampling evaluation of the multi-part contributed some of the shakes as well?",
    "2031293": "My best private score is 0.771 which was used your [binarized data notebook](https://www.kaggle.com/code/takanashihumbert/binarized-data-for-multiome-part) (Great work, thank you so much), but it's a little worse than my selected submission on both CV and public LB, so I didn't select it."
  },
  "source": "meta"
}