{
  "id": 379844,
  "title": "Question mean and std?",
  "url": "/competitions/rsna-breast-cancer-detection/discussion/379844",
  "author_name": "SenTran",
  "post_date": "2023-01-21T08:14:28.417000",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>I see many notebooks setting mean = [0.5, 0.5, 0.5] and std = [0.5, 0.5, 0.5], the setting attual impact on this competition?</p>",
  "messages": [
    {
      "id": 2111430,
      "postDate": "2023-01-22T22:22:24.593Z",
      "content": "<p>As stated by <a href=\"https://www.kaggle.com/aliabdin1\" target=\"_blank\">@aliabdin1</a>, try to compute these statistics on the current dataset. <br>\nThat being said, so far these don't have an important impact on my models.</p>",
      "rawMarkdown": "As stated by @aliabdin1, try to compute these statistics on the current dataset. \nThat being said, so far these don't have an important impact on my models.",
      "votes": 1,
      "replies": [
        {
          "id": 2111636,
          "postDate": "2023-01-23T03:48:54.680Z",
          "content": "<p>Thank you for your answer.</p>",
          "rawMarkdown": "Thank you for your answer."
        }
      ]
    },
    {
      "id": 2111210,
      "postDate": "2023-01-22T17:31:29.400Z",
      "content": "<p>You can calculate these means + stds yourself and try if you can achieve any improvement, see e.g.:<br>\n<a href=\"https://www.kaggle.com/code/aliabdin1/calculate-mean-std-of-images\" target=\"_blank\">https://www.kaggle.com/code/aliabdin1/calculate-mean-std-of-images</a></p>\n<p>Or you can easily test if these settings are useful by taking a normalized image, check if its (channelwise) means are 0 and stds 1 or very close to it.</p>",
      "rawMarkdown": "You can calculate these means + stds yourself and try if you can achieve any improvement, see e.g.:\nhttps://www.kaggle.com/code/aliabdin1/calculate-mean-std-of-images\n\nOr you can easily test if these settings are useful by taking a normalized image, check if its (channelwise) means are 0 and stds 1 or very close to it.",
      "votes": 1,
      "replies": [
        {
          "id": 2111637,
          "postDate": "2023-01-23T03:49:17.623Z",
          "content": "<p>Thank you for your comment.</p>",
          "rawMarkdown": "Thank you for your comment."
        }
      ]
    },
    {
      "id": 2109637,
      "postDate": "2023-01-21T14:54:49.077Z",
      "content": "<p>For me, the setting made my CV worse.</p>",
      "rawMarkdown": "For me, the setting made my CV worse.",
      "votes": 1
    },
    {
      "id": 2109242,
      "postDate": "2023-01-21T08:14:28.417Z",
      "content": "<p>I see many notebooks setting mean = [0.5, 0.5, 0.5] and std = [0.5, 0.5, 0.5], the setting attual impact on this competition?</p>",
      "rawMarkdown": "I see many notebooks setting mean = [0.5, 0.5, 0.5] and std = [0.5, 0.5, 0.5], the setting attual impact on this competition?",
      "votes": 2
    }
  ],
  "comments": [
    {
      "id": 2111430,
      "author_name": "Yassine Alouini",
      "author_url": "",
      "post_date": "2023-01-22T22:22:24.593000",
      "content": "<p>As stated by <a href=\"https://www.kaggle.com/aliabdin1\" target=\"_blank\">@aliabdin1</a>, try to compute these statistics on the current dataset. <br>\nThat being said, so far these don't have an important impact on my models.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 2111636,
          "author_name": "SenTran",
          "author_url": "",
          "post_date": "2023-01-23T03:48:54.680000",
          "content": "<p>Thank you for your answer.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 2111210,
      "author_name": "Ali Abdin",
      "author_url": "",
      "post_date": "2023-01-22T17:31:29.400000",
      "content": "<p>You can calculate these means + stds yourself and try if you can achieve any improvement, see e.g.:<br>\n<a href=\"https://www.kaggle.com/code/aliabdin1/calculate-mean-std-of-images\" target=\"_blank\">https://www.kaggle.com/code/aliabdin1/calculate-mean-std-of-images</a></p>\n<p>Or you can easily test if these settings are useful by taking a normalized image, check if its (channelwise) means are 0 and stds 1 or very close to it.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 2111637,
          "author_name": "SenTran",
          "author_url": "",
          "post_date": "2023-01-23T03:49:17.623000",
          "content": "<p>Thank you for your comment.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 2109637,
      "author_name": "taruto",
      "author_url": "",
      "post_date": "2023-01-21T14:54:49.077000",
      "content": "<p>For me, the setting made my CV worse.</p>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2111430": "As stated by @aliabdin1, try to compute these statistics on the current dataset. \nThat being said, so far these don't have an important impact on my models.",
    "2111210": "You can calculate these means + stds yourself and try if you can achieve any improvement, see e.g.:\nhttps://www.kaggle.com/code/aliabdin1/calculate-mean-std-of-images\n\nOr you can easily test if these settings are useful by taking a normalized image, check if its (channelwise) means are 0 and stds 1 or very close to it.",
    "2109637": "For me, the setting made my CV worse.",
    "2109242": "I see many notebooks setting mean = [0.5, 0.5, 0.5] and std = [0.5, 0.5, 0.5], the setting attual impact on this competition?"
  }
}