{
  "id": 370968,
  "title": "logs when a submission failed",
  "url": "/competitions/rsna-breast-cancer-detection/discussion/370968",
  "author_name": "Laurent Pourchot",
  "post_date": "2022-12-07T10:57:49.764000",
  "votes": 3,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi <a href=\"https://www.kaggle.com/maggiemd\" target=\"_blank\">@maggiemd</a> , is it possible to get the logs when only the final submission failed ? in order to debugg 🤕 when no issues raised during the notebook run. Many thanks</p>",
  "messages": [
    {
      "id": 2057791,
      "postDate": "2022-12-07T10:57:49.763Z",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/maggiemd\" target=\"_blank\">@maggiemd</a> , is it possible to get the logs when only the final submission failed ? in order to debugg 🤕 when no issues raised during the notebook run. Many thanks</p>",
      "rawMarkdown": "Hi @maggiemd , is it possible to get the logs when only the final submission failed ? in order to debugg 🤕 when no issues raised during the notebook run. Many thanks",
      "votes": 3
    },
    {
      "id": 2057994,
      "postDate": "2022-12-07T14:27:26.347Z",
      "content": "<p>The leaderboards already leak a lot of data, imagine what would happen if we got logs.</p>\n<p>One suggestion - try running your inference on all of the train data.  Might help find the bug.</p>\n<p>Also note that there are new machine ids in the pub.</p>",
      "rawMarkdown": "The leaderboards already leak a lot of data, imagine what would happen if we got logs.\n\nOne suggestion - try running your inference on all of the train data.  Might help find the bug.\n\nAlso note that there are new machine ids in the pub.\n\n",
      "votes": 1
    },
    {
      "id": 2057929,
      "postDate": "2022-12-07T13:33:27.283Z",
      "content": "<p>You have to figure out what happaned - part of competition :)<br>\nSecond one (particular in this one) is … optimization - file processing.<br>\nThen 5% to figure out solution.</p>\n<p>Nice to see you Laurent in this competition.</p>",
      "rawMarkdown": "You have to figure out what happaned - part of competition :)\nSecond one (particular in this one) is ... optimization - file processing.\nThen 5% to figure out solution.\n\nNice to see you Laurent in this competition.",
      "votes": 1,
      "replies": [
        {
          "id": 2058044,
          "postDate": "2022-12-07T15:19:39.573Z",
          "content": "<p>Hi all, thank you for your answers. I have still a long way to become just a little efficient 😄 but I won't give up.<br>\n<a href=\"https://www.kaggle.com/Remek\" target=\"_blank\">@Remek</a>, you are the expert on Yolov, did you try segmentation instead ?, I got quick result with it, and this is not the source of my bug 😄 I think I have an issue with the torch dataloader for test and its batch</p>",
          "rawMarkdown": "Hi all, thank you for your answers. I have still a long way to become just a little efficient 😄 but I won't give up.\n@Remek, you are the expert on Yolov, did you try segmentation instead ?, I got quick result with it, and this is not the source of my bug 😄 I think I have an issue with the torch dataloader for test and its batch"
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 2057994,
      "author_name": "@kaggleqrdl",
      "author_url": "",
      "post_date": "2022-12-07T14:27:26.347000",
      "content": "<p>The leaderboards already leak a lot of data, imagine what would happen if we got logs.</p>\n<p>One suggestion - try running your inference on all of the train data.  Might help find the bug.</p>\n<p>Also note that there are new machine ids in the pub.</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 2057929,
      "author_name": "Remek Kinas",
      "author_url": "",
      "post_date": "2022-12-07T13:33:27.283000",
      "content": "<p>You have to figure out what happaned - part of competition :)<br>\nSecond one (particular in this one) is … optimization - file processing.<br>\nThen 5% to figure out solution.</p>\n<p>Nice to see you Laurent in this competition.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 2058044,
          "author_name": "Laurent Pourchot",
          "author_url": "",
          "post_date": "2022-12-07T15:19:39.573000",
          "content": "<p>Hi all, thank you for your answers. I have still a long way to become just a little efficient 😄 but I won't give up.<br>\n<a href=\"https://www.kaggle.com/Remek\" target=\"_blank\">@Remek</a>, you are the expert on Yolov, did you try segmentation instead ?, I got quick result with it, and this is not the source of my bug 😄 I think I have an issue with the torch dataloader for test and its batch</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2057791": "Hi @maggiemd , is it possible to get the logs when only the final submission failed ? in order to debugg 🤕 when no issues raised during the notebook run. Many thanks",
    "2057994": "The leaderboards already leak a lot of data, imagine what would happen if we got logs.\n\nOne suggestion - try running your inference on all of the train data.  Might help find the bug.\n\nAlso note that there are new machine ids in the pub.\n\n",
    "2057929": "You have to figure out what happaned - part of competition :)\nSecond one (particular in this one) is ... optimization - file processing.\nThen 5% to figure out solution.\n\nNice to see you Laurent in this competition."
  }
}