{
  "id": 428290,
  "title": "Huge shake",
  "url": "/competitions/hubmap-hacking-the-human-vasculature/discussion/428290",
  "author_name": "RickyLu",
  "post_date": "2023-08-01T00:14:21.075000",
  "votes": 7,
  "comment_count": 13,
  "views": 0,
  "content": "<p>lb is 0.509 but private score is only 0.315.😭<br>\nMy best private score submission is 0.515 with a lb 0.296.🙃<br>\nWhy there is a huge shake in this competition?</p>",
  "messages": [
    {
      "id": 2367933,
      "postDate": "2023-08-01T00:14:21.077Z",
      "content": "<p>lb is 0.509 but private score is only 0.315.😭<br>\nMy best private score submission is 0.515 with a lb 0.296.🙃<br>\nWhy there is a huge shake in this competition?</p>",
      "rawMarkdown": "lb is 0.509 but private score is only 0.315.😭\nMy best private score submission is 0.515 with a lb 0.296.🙃\nWhy there is a huge shake in this competition?",
      "votes": 7
    },
    {
      "id": 2367959,
      "postDate": "2023-08-01T00:34:40.083Z",
      "content": "<p>Dataset 1 is all you need. My single model with dataset 1 without dialate is Private 0.543, Public 0.51. My best private is 0.597 :(( <img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2832287%2Fe1c7ada4ec5f11c3470b459e7e225c29%2FScreenshot%20from%202023-08-01%2007-15-12.png?generation=1690850101951840&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "Dataset 1 is all you need. My single model with dataset 1 without dialate is Private 0.543, Public 0.51. My best private is 0.597 :(( ![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2832287%2Fe1c7ada4ec5f11c3470b459e7e225c29%2FScreenshot%20from%202023-08-01%2007-15-12.png?generation=1690850101951840&alt=media)",
      "votes": 3,
      "replies": [
        {
          "id": 2367988,
          "postDate": "2023-08-01T00:54:43.330Z",
          "content": "<p>0.597 - so sad :/ <br>\nIs this also only dataset1? </p>",
          "rawMarkdown": "0.597 - so sad :/ \nIs this also only dataset1? ",
          "replies": [
            {
              "id": 2368107,
              "postDate": "2023-08-01T02:36:20.777Z",
              "content": "<p>My mistake is not ensemble using only dataset 1. I ensemble using model train on dataset 1 and model train on dataset 2, and result is unstable.</p>",
              "rawMarkdown": "My mistake is not ensemble using only dataset 1. I ensemble using model train on dataset 1 and model train on dataset 2, and result is unstable."
            }
          ]
        },
        {
          "id": 2368022,
          "postDate": "2023-08-01T01:39:25.827Z",
          "content": "<p>hug hug 😭</p>",
          "rawMarkdown": "hug hug 😭"
        }
      ]
    },
    {
      "id": 2367956,
      "postDate": "2023-08-01T00:30:33.957Z",
      "content": "<p>lb0.385—private0.461，lb0.53—private0.351。For me maybe the reasion is pseudo labels.</p>",
      "rawMarkdown": "lb0.385—private0.461，lb0.53—private0.351。For me maybe the reasion is pseudo labels.",
      "replies": [
        {
          "id": 2368032,
          "postDate": "2023-08-01T01:43:38.607Z",
          "content": "<p>d2 with sparse annotation is a trap😑</p>",
          "rawMarkdown": "d2 with sparse annotation is a trap😑",
          "replies": [
            {
              "id": 2368168,
              "postDate": "2023-08-01T03:45:55.467Z",
              "content": "<p>看来粗标注，是字面意义上的粗呀😭</p>",
              "rawMarkdown": "看来粗标注，是字面意义上的粗呀😭",
              "votes": 1
            }
          ]
        },
        {
          "id": 2368167,
          "postDate": "2023-08-01T03:45:19.867Z",
          "content": "<p>I tried a late submission without dilate, then lb0.53—private0.351 became lb0.456—private0.512.  So dilate is the main reason…</p>",
          "rawMarkdown": "I tried a late submission without dilate, then lb0.53—private0.351 became lb0.456—private0.512.  So dilate is the main reason..."
        }
      ]
    },
    {
      "id": 2367951,
      "postDate": "2023-08-01T00:25:23.120Z",
      "content": "<p>Another sad one here. My best private score is my best emsemble submission. I didn't trust him. He punished me.</p>",
      "rawMarkdown": "Another sad one here. My best private score is my best emsemble submission. I didn't trust him. He punished me.",
      "replies": [
        {
          "id": 2368038,
          "postDate": "2023-08-01T01:44:41.203Z",
          "rawMarkdown": "",
          "isDeleted": true
        },
        {
          "id": 2368039,
          "postDate": "2023-08-01T01:45:18.677Z",
          "content": "<p>lb is an illusion in this comp🫠</p>",
          "rawMarkdown": "lb is an illusion in this comp🫠"
        }
      ]
    },
    {
      "id": 2368058,
      "postDate": "2023-08-01T02:00:25.073Z",
      "content": "<p>Just another sad one. For me is the dilation. Not working!</p>",
      "rawMarkdown": "Just another sad one. For me is the dilation. Not working!",
      "isDeleted": true,
      "replies": [
        {
          "id": 2368099,
          "postDate": "2023-08-01T02:30:18.613Z",
          "content": "<p>dilation is a \"shuangrenjian\"🤣</p>",
          "rawMarkdown": "dilation is a \"shuangrenjian\"🤣"
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 2367959,
      "author_name": "Quan Vu",
      "author_url": "",
      "post_date": "2023-08-01T00:34:40.083000",
      "content": "<p>Dataset 1 is all you need. My single model with dataset 1 without dialate is Private 0.543, Public 0.51. My best private is 0.597 :(( <img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2832287%2Fe1c7ada4ec5f11c3470b459e7e225c29%2FScreenshot%20from%202023-08-01%2007-15-12.png?generation=1690850101951840&amp;alt=media\" alt=\"\"></p>",
      "votes": 3,
      "replies": [
        {
          "id": 2367988,
          "author_name": "tsobolev",
          "author_url": "",
          "post_date": "2023-08-01T00:54:43.330000",
          "content": "<p>0.597 - so sad :/ <br>\nIs this also only dataset1? </p>",
          "votes": 0,
          "replies": [
            {
              "id": 2368107,
              "author_name": "Quan Vu",
              "author_url": "",
              "post_date": "2023-08-01T02:36:20.777000",
              "content": "<p>My mistake is not ensemble using only dataset 1. I ensemble using model train on dataset 1 and model train on dataset 2, and result is unstable.</p>",
              "votes": 0,
              "replies": []
            }
          ]
        },
        {
          "id": 2368022,
          "author_name": "RickyLu",
          "author_url": "",
          "post_date": "2023-08-01T01:39:25.827000",
          "content": "<p>hug hug 😭</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 2367956,
      "author_name": "FreddyHu",
      "author_url": "",
      "post_date": "2023-08-01T00:30:33.957000",
      "content": "<p>lb0.385—private0.461，lb0.53—private0.351。For me maybe the reasion is pseudo labels.</p>",
      "votes": 0,
      "replies": [
        {
          "id": 2368032,
          "author_name": "RickyLu",
          "author_url": "",
          "post_date": "2023-08-01T01:43:38.607000",
          "content": "<p>d2 with sparse annotation is a trap😑</p>",
          "votes": 0,
          "replies": [
            {
              "id": 2368168,
              "author_name": "FreddyHu",
              "author_url": "",
              "post_date": "2023-08-01T03:45:55.467000",
              "content": "<p>看来粗标注，是字面意义上的粗呀😭</p>",
              "votes": 1,
              "replies": []
            }
          ]
        },
        {
          "id": 2368167,
          "author_name": "FreddyHu",
          "author_url": "",
          "post_date": "2023-08-01T03:45:19.867000",
          "content": "<p>I tried a late submission without dilate, then lb0.53—private0.351 became lb0.456—private0.512.  So dilate is the main reason…</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 2367951,
      "author_name": "阳光开朗大男孩",
      "author_url": "",
      "post_date": "2023-08-01T00:25:23.120000",
      "content": "<p>Another sad one here. My best private score is my best emsemble submission. I didn't trust him. He punished me.</p>",
      "votes": 0,
      "replies": [
        {
          "id": 2368038,
          "author_name": "",
          "author_url": "",
          "post_date": "2023-08-01T01:44:41.203000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 2368039,
          "author_name": "RickyLu",
          "author_url": "",
          "post_date": "2023-08-01T01:45:18.677000",
          "content": "<p>lb is an illusion in this comp🫠</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 2368058,
      "author_name": "",
      "author_url": "",
      "post_date": "2023-08-01T02:00:25.073000",
      "content": "<p>Just another sad one. For me is the dilation. Not working!</p>",
      "votes": 0,
      "replies": [
        {
          "id": 2368099,
          "author_name": "RickyLu",
          "author_url": "",
          "post_date": "2023-08-01T02:30:18.613000",
          "content": "<p>dilation is a \"shuangrenjian\"🤣</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2367933": "lb is 0.509 but private score is only 0.315.😭\nMy best private score submission is 0.515 with a lb 0.296.🙃\nWhy there is a huge shake in this competition?",
    "2367959": "Dataset 1 is all you need. My single model with dataset 1 without dialate is Private 0.543, Public 0.51. My best private is 0.597 :(( ![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2832287%2Fe1c7ada4ec5f11c3470b459e7e225c29%2FScreenshot%20from%202023-08-01%2007-15-12.png?generation=1690850101951840&alt=media)",
    "2367956": "lb0.385—private0.461，lb0.53—private0.351。For me maybe the reasion is pseudo labels.",
    "2367951": "Another sad one here. My best private score is my best emsemble submission. I didn't trust him. He punished me.",
    "2368058": "Just another sad one. For me is the dilation. Not working!"
  }
}