{
  "id": 422204,
  "title": "Submission Score zero??",
  "url": "/competitions/hubmap-hacking-the-human-vasculature/discussion/422204",
  "author_name": "",
  "post_date": "2023-07-08T18:04:11.753439400Z",
  "votes": 4,
  "comment_count": 12,
  "views": 0,
  "content": "<p>I recently submitted a model but gave me 0 score after submission.<br>\nThe submission format was the same…</p>\n<p>any one else facing the same issue??</p>\n<p>why is this hapenning??</p>",
  "messages": [
    {
      "id": "2335686",
      "postDate": "07/08/2023 18:04:11",
      "content": "<p>I recently submitted a model but gave me 0 score after submission.<br>\nThe submission format was the same…</p>\n<p>any one else facing the same issue??</p>\n<p>why is this hapenning??</p>",
      "rawMarkdown": "I recently submitted a model but gave me 0 score after submission.\nThe submission format was the same...\n\nany one else facing the same issue??\n\nwhy is this hapenning??",
      "votes": null
    },
    {
      "id": "2335706",
      "postDate": "07/08/2023 18:17:06",
      "content": "<p>yes, i had two submissions with zero</p>",
      "rawMarkdown": "yes, i had two submissions with zero",
      "votes": null
    },
    {
      "id": "2335724",
      "postDate": "07/08/2023 18:31:33",
      "content": "<p>It is possible that your prediction is empty.</p>",
      "rawMarkdown": "It is possible that your prediction is empty.",
      "votes": null
    },
    {
      "id": "2336197",
      "postDate": "07/09/2023 08:01:52",
      "content": "<p>In my case predicted width and height were not 512*512</p>",
      "rawMarkdown": "In my case predicted width and height were not 512*512",
      "votes": null
    },
    {
      "id": "2336492",
      "postDate": "07/09/2023 12:02:35",
      "content": "<p>you mean in the hidden data??</p>",
      "rawMarkdown": "you mean in the hidden data??",
      "votes": null
    },
    {
      "id": "2336495",
      "postDate": "07/09/2023 12:03:09",
      "content": "<p>Thanks for the input 👍👍</p>",
      "rawMarkdown": "Thanks for the input 👍👍",
      "votes": null
    },
    {
      "id": "2338299",
      "postDate": "07/10/2023 17:09:03",
      "content": "<p>Hello, I am facing the same issue. I have received a score of zero in all four attempts. I am not getting submission error. In my first attempt, I submitted different instances under a single mask, but then I realized that I need to submit each instance with a separate mask, yet I still got zero. I assumed all test images were of size 512x512 and filled the ‘height’ and ‘width’ columns in the submission dataframe with 512, but I corrected that as well and still received zero score. There is also the output generated by the model for the provided test image. I don’t think the model is generating empty images. What should I pay attention to?</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3212617%2F767bece10d0ce9fd3d882747d7c430a7%2FScreenshot%202023-07-10%20at%2020.07.52.png?generation=1689008929305494&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "Hello, I am facing the same issue. I have received a score of zero in all four attempts. I am not getting submission error. In my first attempt, I submitted different instances under a single mask, but then I realized that I need to submit each instance with a separate mask, yet I still got zero. I assumed all test images were of size 512x512 and filled the ‘height’ and ‘width’ columns in the submission dataframe with 512, but I corrected that as well and still received zero score. There is also the output generated by the model for the provided test image. I don’t think the model is generating empty images. What should I pay attention to?\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3212617%2F767bece10d0ce9fd3d882747d7c430a7%2FScreenshot%202023-07-10%20at%2020.07.52.png?generation=1689008929305494&alt=media)",
      "votes": null
    },
    {
      "id": "2340961",
      "postDate": "07/11/2023 19:26:57",
      "content": "<p>Did you figure something out?</p>",
      "rawMarkdown": "Did you figure something out?",
      "votes": null
    },
    {
      "id": "2340962",
      "postDate": "07/11/2023 19:27:05",
      "content": "<p>Same problem</p>",
      "rawMarkdown": "Same problem",
      "votes": null
    },
    {
      "id": "2341536",
      "postDate": "07/12/2023 06:47:51",
      "content": "<p>Yeah same issue I am also facing.<br>\nAnyone knows what to be done ??</p>",
      "rawMarkdown": "Yeah same issue I am also facing.\nAnyone knows what to be done ??",
      "votes": null
    },
    {
      "id": "2342033",
      "postDate": "07/12/2023 12:59:35",
      "content": "<p>Solved. I was using skimage binary dilation for seperating instances after predict semantic segmentation mask, but while I was seperating instances, instance segmentation masks consist of only contours, it did not fill the holes so I got zero score. My score is terrible but at least it is not zero. </p>",
      "rawMarkdown": "Solved. I was using skimage binary dilation for seperating instances after predict semantic segmentation mask, but while I was seperating instances, instance segmentation masks consist of only contours, it did not fill the holes so I got zero score. My score is terrible but at least it is not zero.",
      "votes": null
    },
    {
      "id": "2342931",
      "postDate": "07/13/2023 09:08:28",
      "content": "<p>Do you slove this problem? I also meet this problem.</p>",
      "rawMarkdown": "Do you slove this problem? I also meet this problem.",
      "votes": null
    },
    {
      "id": "2342932",
      "postDate": "07/13/2023 09:08:56",
      "content": "<p>Do you slove it? </p>",
      "rawMarkdown": "Do you slove it?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2335706,
      "author_name": "vaibhavpatel11",
      "author_url": "",
      "post_date": "07/08/2023 18:17:06",
      "content": "<p>yes, i had two submissions with zero</p>",
      "votes": null,
      "replies": [
        {
          "id": 2340961,
          "author_name": "satwato",
          "author_url": "",
          "post_date": "07/11/2023 19:26:57",
          "content": "<p>Did you figure something out?</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2335724,
      "author_name": "bent1e",
      "author_url": "",
      "post_date": "07/08/2023 18:31:33",
      "content": "<p>It is possible that your prediction is empty.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2336492,
          "author_name": "vishakkbhat",
          "author_url": "",
          "post_date": "07/09/2023 12:02:35",
          "content": "<p>you mean in the hidden data??</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2336197,
      "author_name": "dodoit",
      "author_url": "",
      "post_date": "07/09/2023 08:01:52",
      "content": "<p>In my case predicted width and height were not 512*512</p>",
      "votes": null,
      "replies": [
        {
          "id": 2336495,
          "author_name": "vishakkbhat",
          "author_url": "",
          "post_date": "07/09/2023 12:03:09",
          "content": "<p>Thanks for the input 👍👍</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2338299,
      "author_name": "turkenm",
      "author_url": "",
      "post_date": "07/10/2023 17:09:03",
      "content": "<p>Hello, I am facing the same issue. I have received a score of zero in all four attempts. I am not getting submission error. In my first attempt, I submitted different instances under a single mask, but then I realized that I need to submit each instance with a separate mask, yet I still got zero. I assumed all test images were of size 512x512 and filled the ‘height’ and ‘width’ columns in the submission dataframe with 512, but I corrected that as well and still received zero score. There is also the output generated by the model for the provided test image. I don’t think the model is generating empty images. What should I pay attention to?</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3212617%2F767bece10d0ce9fd3d882747d7c430a7%2FScreenshot%202023-07-10%20at%2020.07.52.png?generation=1689008929305494&amp;alt=media\" alt=\"\"></p>",
      "votes": null,
      "replies": [
        {
          "id": 2341536,
          "author_name": "vishakkbhat",
          "author_url": "",
          "post_date": "07/12/2023 06:47:51",
          "content": "<p>Yeah same issue I am also facing.<br>\nAnyone knows what to be done ??</p>",
          "votes": null,
          "replies": [
            {
              "id": 2342033,
              "author_name": "turkenm",
              "author_url": "",
              "post_date": "07/12/2023 12:59:35",
              "content": "<p>Solved. I was using skimage binary dilation for seperating instances after predict semantic segmentation mask, but while I was seperating instances, instance segmentation masks consist of only contours, it did not fill the holes so I got zero score. My score is terrible but at least it is not zero. </p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 2340962,
      "author_name": "satwato",
      "author_url": "",
      "post_date": "07/11/2023 19:27:05",
      "content": "<p>Same problem</p>",
      "votes": null,
      "replies": [
        {
          "id": 2342932,
          "author_name": "xiaoxuxu",
          "author_url": "",
          "post_date": "07/13/2023 09:08:56",
          "content": "<p>Do you slove it? </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2342931,
      "author_name": "xiaoxuxu",
      "author_url": "",
      "post_date": "07/13/2023 09:08:28",
      "content": "<p>Do you slove this problem? I also meet this problem.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2335686": "I recently submitted a model but gave me 0 score after submission.\nThe submission format was the same...\n\nany one else facing the same issue??\n\nwhy is this hapenning??",
    "2335706": "yes, i had two submissions with zero",
    "2335724": "It is possible that your prediction is empty.",
    "2336197": "In my case predicted width and height were not 512*512",
    "2336492": "you mean in the hidden data??",
    "2336495": "Thanks for the input 👍👍",
    "2338299": "Hello, I am facing the same issue. I have received a score of zero in all four attempts. I am not getting submission error. In my first attempt, I submitted different instances under a single mask, but then I realized that I need to submit each instance with a separate mask, yet I still got zero. I assumed all test images were of size 512x512 and filled the ‘height’ and ‘width’ columns in the submission dataframe with 512, but I corrected that as well and still received zero score. There is also the output generated by the model for the provided test image. I don’t think the model is generating empty images. What should I pay attention to?\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3212617%2F767bece10d0ce9fd3d882747d7c430a7%2FScreenshot%202023-07-10%20at%2020.07.52.png?generation=1689008929305494&alt=media)",
    "2340961": "Did you figure something out?",
    "2340962": "Same problem",
    "2341536": "Yeah same issue I am also facing.\nAnyone knows what to be done ??",
    "2342033": "Solved. I was using skimage binary dilation for seperating instances after predict semantic segmentation mask, but while I was seperating instances, instance segmentation masks consist of only contours, it did not fill the holes so I got zero score. My score is terrible but at least it is not zero.",
    "2342931": "Do you slove this problem? I also meet this problem.",
    "2342932": "Do you slove it?"
  },
  "source": "meta"
}