{
  "id": 416186,
  "title": "Submission failed due to Notebook Threw Exception",
  "url": "/competitions/benetech-making-graphs-accessible/discussion/416186",
  "author_name": "",
  "post_date": "2023-06-10T04:00:38.147303500Z",
  "votes": null,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hii everyone, </p>\n<p>I am submitting my notebook but it's throwing me this error. The notebook is running on the given public dataset but not running after that. I am not sure what can be done as I am new to this type of code based (notebook submission) competition. Any help will be highly appreciated. Thanks!</p>",
  "messages": [
    {
      "id": "2294408",
      "postDate": "06/10/2023 04:00:38",
      "content": "<p>Hii everyone, </p>\n<p>I am submitting my notebook but it's throwing me this error. The notebook is running on the given public dataset but not running after that. I am not sure what can be done as I am new to this type of code based (notebook submission) competition. Any help will be highly appreciated. Thanks!</p>",
      "rawMarkdown": "Hii everyone, \n\nI am submitting my notebook but it's throwing me this error. The notebook is running on the given public dataset but not running after that. I am not sure what can be done as I am new to this type of code based (notebook submission) competition. Any help will be highly appreciated. Thanks!",
      "votes": null
    },
    {
      "id": "2294544",
      "postDate": "06/10/2023 06:32:26",
      "content": "<p>There can be many reasons, but it's difficult to point ot without the code. Some suggestions u might try:</p>\n<ol>\n<li>Public test data is just a placeholder. It gets replaced by 4000 test data. Your algo should be able to process them all.</li>\n<li>Some images in test data may be greyscale. Your processing loop should be able to handle that as well.</li>\n<li>Your code should be able to handle Nan outputs.</li>\n</ol>\n<p>There are many inference notebooks in the code tab of this competition. Check them out.</p>",
      "rawMarkdown": "There can be many reasons, but it's difficult to point ot without the code. Some suggestions u might try:\n\n1. Public test data is just a placeholder. It gets replaced by 4000 test data. Your algo should be able to process them all.\n2. Some images in test data may be greyscale. Your processing loop should be able to handle that as well.\n3. Your code should be able to handle Nan outputs.\n\n\nThere are many inference notebooks in the code tab of this competition. Check them out.",
      "votes": null
    },
    {
      "id": "2307691",
      "postDate": "06/18/2023 11:16:57",
      "content": "<p><a href=\"https://www.kaggle.com/sandeepmnair\" target=\"_blank\">@sandeepmnair</a> I am running through the same problem. How do you know that some images are greyscale? also nan outputs? I cheched the code tab and I did not find any inference notebook. Thanks for your reply.</p>",
      "rawMarkdown": "sandeepmnair I am running through the same problem. How do you know that some images are greyscale? also nan outputs? I cheched the code tab and I did not find any inference notebook. Thanks for your reply.",
      "votes": null
    },
    {
      "id": "2307693",
      "postDate": "06/18/2023 11:18:10",
      "content": "<p>Have you discovered the error where it is? I am running the same situation.</p>",
      "rawMarkdown": "Have you discovered the error where it is? I am running the same situation.",
      "votes": null
    },
    {
      "id": "2308007",
      "postDate": "06/18/2023 15:34:52",
      "content": "<p>When I trained Matcha with max length 512 and submitted to inference more than the trained length (such as 768, 1024), it threw an Exception. But same model trained with more epochs didn't caused such error.<br>\nI don't know the exact reason but I think the improperly trained model generates unexpected output.</p>",
      "rawMarkdown": "When I trained Matcha with max length 512 and submitted to inference more than the trained length (such as 768, 1024), it threw an Exception. But same model trained with more epochs didn't caused such error.\nI don't know the exact reason but I think the improperly trained model generates unexpected output.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2294544,
      "author_name": "sandeepmnair",
      "author_url": "",
      "post_date": "06/10/2023 06:32:26",
      "content": "<p>There can be many reasons, but it's difficult to point ot without the code. Some suggestions u might try:</p>\n<ol>\n<li>Public test data is just a placeholder. It gets replaced by 4000 test data. Your algo should be able to process them all.</li>\n<li>Some images in test data may be greyscale. Your processing loop should be able to handle that as well.</li>\n<li>Your code should be able to handle Nan outputs.</li>\n</ol>\n<p>There are many inference notebooks in the code tab of this competition. Check them out.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2307691,
          "author_name": "hamzahabduljalil",
          "author_url": "",
          "post_date": "06/18/2023 11:16:57",
          "content": "<p><a href=\"https://www.kaggle.com/sandeepmnair\" target=\"_blank\">@sandeepmnair</a> I am running through the same problem. How do you know that some images are greyscale? also nan outputs? I cheched the code tab and I did not find any inference notebook. Thanks for your reply.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2307693,
      "author_name": "hamzahabduljalil",
      "author_url": "",
      "post_date": "06/18/2023 11:18:10",
      "content": "<p>Have you discovered the error where it is? I am running the same situation.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2308007,
      "author_name": "dodoit",
      "author_url": "",
      "post_date": "06/18/2023 15:34:52",
      "content": "<p>When I trained Matcha with max length 512 and submitted to inference more than the trained length (such as 768, 1024), it threw an Exception. But same model trained with more epochs didn't caused such error.<br>\nI don't know the exact reason but I think the improperly trained model generates unexpected output.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2294408": "Hii everyone, \n\nI am submitting my notebook but it's throwing me this error. The notebook is running on the given public dataset but not running after that. I am not sure what can be done as I am new to this type of code based (notebook submission) competition. Any help will be highly appreciated. Thanks!",
    "2294544": "There can be many reasons, but it's difficult to point ot without the code. Some suggestions u might try:\n\n1. Public test data is just a placeholder. It gets replaced by 4000 test data. Your algo should be able to process them all.\n2. Some images in test data may be greyscale. Your processing loop should be able to handle that as well.\n3. Your code should be able to handle Nan outputs.\n\n\nThere are many inference notebooks in the code tab of this competition. Check them out.",
    "2307691": "sandeepmnair I am running through the same problem. How do you know that some images are greyscale? also nan outputs? I cheched the code tab and I did not find any inference notebook. Thanks for your reply.",
    "2307693": "Have you discovered the error where it is? I am running the same situation.",
    "2308007": "When I trained Matcha with max length 512 and submitted to inference more than the trained length (such as 768, 1024), it threw an Exception. But same model trained with more epochs didn't caused such error.\nI don't know the exact reason but I think the improperly trained model generates unexpected output."
  },
  "source": "meta"
}