{
  "id": 319335,
  "title": "I can't wait to see how top players crossed 0.88",
  "url": "/competitions/happy-whale-and-dolphin/discussion/319335",
  "author_name": "",
  "post_date": "2022-04-16T17:33:34.089714400Z",
  "votes": 7,
  "comment_count": 11,
  "views": 0,
  "content": "<p>I can't wait to see how the top players scored over 0.88. I tried quite a bit, but I didn't even get over 0.86. I'm so curious about how the puzzles are solved, and I hope the competition is over when I wake up.</p>",
  "messages": [
    {
      "id": "1757506",
      "postDate": "04/16/2022 17:33:34",
      "content": "<p>I can't wait to see how the top players scored over 0.88. I tried quite a bit, but I didn't even get over 0.86. I'm so curious about how the puzzles are solved, and I hope the competition is over when I wake up.</p>",
      "rawMarkdown": "I can't wait to see how the top players scored over 0.88. I tried quite a bit, but I didn't even get over 0.86. I'm so curious about how the puzzles are solved, and I hope the competition is over when I wake up.",
      "votes": null
    },
    {
      "id": "1757752",
      "postDate": "04/17/2022 01:34:38",
      "content": "<p>I guess proper post-processing really boosts the score :D </p>",
      "rawMarkdown": "I guess proper post-processing really boosts the score :D",
      "votes": null
    },
    {
      "id": "1757855",
      "postDate": "04/17/2022 04:38:39",
      "content": "<p>i think so. i have no idea about it totally!!</p>",
      "rawMarkdown": "i think so. i have no idea about it totally!!",
      "votes": null
    },
    {
      "id": "1757961",
      "postDate": "04/17/2022 07:06:48",
      "content": "<p>It depends on how much time you spend  and how good your algorithm and hardware are.</p>\n<p>maybe just perfectly fitted for public test dataset.</p>",
      "rawMarkdown": "It depends on how much time you spend  and how good your algorithm and hardware are.\n\nmaybe just perfectly fitted for public test dataset.",
      "votes": null
    },
    {
      "id": "1758530",
      "postDate": "04/17/2022 18:47:00",
      "content": "<p>Wow, LB scores are still increasing. Looks like the top teams are above 0.89 now and Gold is above 0.88. Great job everyone! I'm looking forward to learning from the top solutions!</p>",
      "rawMarkdown": "Wow, LB scores are still increasing. Looks like the top teams are above 0.89 now and Gold is above 0.88. Great job everyone! I'm looking forward to learning from the top solutions!",
      "votes": null
    },
    {
      "id": "1758598",
      "postDate": "04/17/2022 20:33:23",
      "content": "<p>I'm hopes are hanging on that thought Dragon.</p>",
      "rawMarkdown": "I'm hopes are hanging on that thought Dragon.",
      "votes": null
    },
    {
      "id": "1758681",
      "postDate": "04/18/2022 00:07:59",
      "content": "<p>maybe leverage different models rather than just efficient net</p>",
      "rawMarkdown": "maybe leverage different models rather than just efficient net",
      "votes": null
    },
    {
      "id": "1758771",
      "postDate": "04/18/2022 03:45:41",
      "content": "<p>My team score without post processing and didn't get over 0.87.</p>",
      "rawMarkdown": "My team score without post processing and didn't get over 0.87.",
      "votes": null
    },
    {
      "id": "1758797",
      "postDate": "04/18/2022 04:31:38",
      "content": "<p>I didn't do post processing, too. we don't have much time to investigate data. we can get 0.858 PUBLIC LB<br>\nWe just concatenated embeddings and did KNN. set new_individuals by a specific threshold<br>\nHow can we fit new_individuals? i don't think it makes sense. or maybe I am missing something</p>",
      "rawMarkdown": "I didn't do post processing, too. we don't have much time to investigate data. we can get 0.858 PUBLIC LB\nWe just concatenated embeddings and did KNN. set new_individuals by a specific threshold\nHow can we fit new_individuals? i don't think it makes sense. or maybe I am missing something",
      "votes": null
    },
    {
      "id": "1758877",
      "postDate": "04/18/2022 05:56:06",
      "content": "<p>It's interesting to see how the top players are doing. I'm curious to see if Gold can maintain their lead or if someone will surpass them.</p>",
      "rawMarkdown": "It's interesting to see how the top players are doing. I'm curious to see if Gold can maintain their lead or if someone will surpass them.",
      "votes": null
    },
    {
      "id": "1758997",
      "postDate": "04/18/2022 08:42:12",
      "content": "<p>Its incredible to see drastic changes in the last two days , getting over 0.89 😲 , I think it has got something to do with new individuals prediction</p>",
      "rawMarkdown": "Its incredible to see drastic changes in the last two days , getting over 0.89 😲 , I think it has got something to do with new individuals prediction",
      "votes": null
    },
    {
      "id": "1759106",
      "postDate": "04/18/2022 10:44:18",
      "content": "<p>I'm also looking forward to learning from the top solutions， especially the tricks we missed in this competition.</p>",
      "rawMarkdown": "I'm also looking forward to learning from the top solutions， especially the tricks we missed in this competition.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1757752,
      "author_name": "ptran1203",
      "author_url": "",
      "post_date": "04/17/2022 01:34:38",
      "content": "<p>I guess proper post-processing really boosts the score :D </p>",
      "votes": null,
      "replies": [
        {
          "id": 1757855,
          "author_name": "deepkim",
          "author_url": "",
          "post_date": "04/17/2022 04:38:39",
          "content": "<p>i think so. i have no idea about it totally!!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1758681,
          "author_name": "liuzhangzhen",
          "author_url": "",
          "post_date": "04/18/2022 00:07:59",
          "content": "<p>maybe leverage different models rather than just efficient net</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1757961,
      "author_name": "dragonzhang",
      "author_url": "",
      "post_date": "04/17/2022 07:06:48",
      "content": "<p>It depends on how much time you spend  and how good your algorithm and hardware are.</p>\n<p>maybe just perfectly fitted for public test dataset.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1758598,
          "author_name": "mutantspore",
          "author_url": "",
          "post_date": "04/17/2022 20:33:23",
          "content": "<p>I'm hopes are hanging on that thought Dragon.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1758530,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "04/17/2022 18:47:00",
      "content": "<p>Wow, LB scores are still increasing. Looks like the top teams are above 0.89 now and Gold is above 0.88. Great job everyone! I'm looking forward to learning from the top solutions!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1759106,
          "author_name": "librauee",
          "author_url": "",
          "post_date": "04/18/2022 10:44:18",
          "content": "<p>I'm also looking forward to learning from the top solutions， especially the tricks we missed in this competition.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1758771,
      "author_name": "duykhanh99",
      "author_url": "",
      "post_date": "04/18/2022 03:45:41",
      "content": "<p>My team score without post processing and didn't get over 0.87.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1758797,
          "author_name": "deepkim",
          "author_url": "",
          "post_date": "04/18/2022 04:31:38",
          "content": "<p>I didn't do post processing, too. we don't have much time to investigate data. we can get 0.858 PUBLIC LB<br>\nWe just concatenated embeddings and did KNN. set new_individuals by a specific threshold<br>\nHow can we fit new_individuals? i don't think it makes sense. or maybe I am missing something</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1758877,
      "author_name": "",
      "author_url": "",
      "post_date": "04/18/2022 05:56:06",
      "content": "<p>It's interesting to see how the top players are doing. I'm curious to see if Gold can maintain their lead or if someone will surpass them.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1758997,
      "author_name": "tanulsingh077",
      "author_url": "",
      "post_date": "04/18/2022 08:42:12",
      "content": "<p>Its incredible to see drastic changes in the last two days , getting over 0.89 😲 , I think it has got something to do with new individuals prediction</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1757506": "I can't wait to see how the top players scored over 0.88. I tried quite a bit, but I didn't even get over 0.86. I'm so curious about how the puzzles are solved, and I hope the competition is over when I wake up.",
    "1757752": "I guess proper post-processing really boosts the score :D",
    "1757855": "i think so. i have no idea about it totally!!",
    "1757961": "It depends on how much time you spend  and how good your algorithm and hardware are.\n\nmaybe just perfectly fitted for public test dataset.",
    "1758530": "Wow, LB scores are still increasing. Looks like the top teams are above 0.89 now and Gold is above 0.88. Great job everyone! I'm looking forward to learning from the top solutions!",
    "1758598": "I'm hopes are hanging on that thought Dragon.",
    "1758681": "maybe leverage different models rather than just efficient net",
    "1758771": "My team score without post processing and didn't get over 0.87.",
    "1758797": "I didn't do post processing, too. we don't have much time to investigate data. we can get 0.858 PUBLIC LB\nWe just concatenated embeddings and did KNN. set new_individuals by a specific threshold\nHow can we fit new_individuals? i don't think it makes sense. or maybe I am missing something",
    "1758877": "It's interesting to see how the top players are doing. I'm curious to see if Gold can maintain their lead or if someone will surpass them.",
    "1758997": "Its incredible to see drastic changes in the last two days , getting over 0.89 😲 , I think it has got something to do with new individuals prediction",
    "1759106": "I'm also looking forward to learning from the top solutions， especially the tricks we missed in this competition."
  },
  "source": "meta"
}