{
  "id": 187460,
  "title": "Submission File not Found",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/187460",
  "author_name": "",
  "post_date": "2020-09-29T04:12:01.335307600Z",
  "votes": 6,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Anyone here got the submission error \"Submission file not found\"? In my case, when I run in \"local\" notebook, there is a submission file at \"output\" directory listing but when I do \"save and Run all (Commit)\" and try to submit, it says \" Submission file not found\". <br>\nThanks in advances.</p>",
  "messages": [
    {
      "id": "1030867",
      "postDate": "09/29/2020 04:12:01",
      "content": "<p>Anyone here got the submission error \"Submission file not found\"? In my case, when I run in \"local\" notebook, there is a submission file at \"output\" directory listing but when I do \"save and Run all (Commit)\" and try to submit, it says \" Submission file not found\". <br>\nThanks in advances.</p>",
      "rawMarkdown": "Anyone here got the submission error \"Submission file not found\"? In my case, when I run in \"local\" notebook, there is a submission file at \"output\" directory listing but when I do \"save and Run all (Commit)\" and try to submit, it says \" Submission file not found\". \nThanks in advances.",
      "votes": null
    },
    {
      "id": "1034703",
      "postDate": "10/02/2020 05:20:07",
      "content": "<p>Hey <a href=\"https://www.kaggle.com/tejshahi\" target=\"_blank\">@tejshahi</a>, I had the same issue a while ago. Confused, I sought out to find the source of source of this error. It helped to separate the model predictions and writing to a submission file into separate code snippets. To debug I did this: </p>\n<ol>\n<li>Initialized the FVC values to 0 and confidence to 250. (Cell block 1)</li>\n<li>Made the model predictions. (Cell block 2)</li>\n<li>Saved them to submission file. (Cell block 3)</li>\n</ol>\n<p>To my surprise, the submission scored <code>-11.1059</code> which is exact score one would get if the model predictions failed. </p>\n<p><strong>The Fix:</strong> Removing a multi-line comment helped me fix the problematic code block. Multiline comments inside code blocks seemed to throw some exception during the notebook execution on the private test run (Works fine during commit though).</p>\n<p>Similar issues for other people were resolved by:</p>\n<ul>\n<li>Removing all files in the working directory, but for the submission.csv file.</li>\n<li>Removing empty code cells and markdown cells.</li>\n<li>Reading certain .dcm files: Try putting them inside a try-except block, make submission in seperate code block and see how it scores. Does it score near to your CV or does it score abnormally?</li>\n</ul>\n<p>Hope this helps :)</p>",
      "rawMarkdown": "Hey @tejshahi, I had the same issue a while ago. Confused, I sought out to find the source of source of this error. It helped to separate the model predictions and writing to a submission file into separate code snippets. To debug I did this: \n 1. Initialized the FVC values to 0 and confidence to 250. (Cell block 1)\n 2. Made the model predictions. (Cell block 2)\n 3. Saved them to submission file. (Cell block 3)\n\nTo my surprise, the submission scored `-11.1059` which is exact score one would get if the model predictions failed. \n\n**The Fix:** Removing a multi-line comment helped me fix the problematic code block. Multiline comments inside code blocks seemed to throw some exception during the notebook execution on the private test run (Works fine during commit though).\n\nSimilar issues for other people were resolved by:\n- Removing all files in the working directory, but for the submission.csv file.\n- Removing empty code cells and markdown cells.\n- Reading certain .dcm files: Try putting them inside a try-except block, make submission in seperate code block and see how it scores. Does it score near to your CV or does it score abnormally?\n\nHope this helps :)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1034703,
      "author_name": "doctorkael",
      "author_url": "",
      "post_date": "10/02/2020 05:20:07",
      "content": "<p>Hey <a href=\"https://www.kaggle.com/tejshahi\" target=\"_blank\">@tejshahi</a>, I had the same issue a while ago. Confused, I sought out to find the source of source of this error. It helped to separate the model predictions and writing to a submission file into separate code snippets. To debug I did this: </p>\n<ol>\n<li>Initialized the FVC values to 0 and confidence to 250. (Cell block 1)</li>\n<li>Made the model predictions. (Cell block 2)</li>\n<li>Saved them to submission file. (Cell block 3)</li>\n</ol>\n<p>To my surprise, the submission scored <code>-11.1059</code> which is exact score one would get if the model predictions failed. </p>\n<p><strong>The Fix:</strong> Removing a multi-line comment helped me fix the problematic code block. Multiline comments inside code blocks seemed to throw some exception during the notebook execution on the private test run (Works fine during commit though).</p>\n<p>Similar issues for other people were resolved by:</p>\n<ul>\n<li>Removing all files in the working directory, but for the submission.csv file.</li>\n<li>Removing empty code cells and markdown cells.</li>\n<li>Reading certain .dcm files: Try putting them inside a try-except block, make submission in seperate code block and see how it scores. Does it score near to your CV or does it score abnormally?</li>\n</ul>\n<p>Hope this helps :)</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1030867": "Anyone here got the submission error \"Submission file not found\"? In my case, when I run in \"local\" notebook, there is a submission file at \"output\" directory listing but when I do \"save and Run all (Commit)\" and try to submit, it says \" Submission file not found\". \nThanks in advances.",
    "1034703": "Hey @tejshahi, I had the same issue a while ago. Confused, I sought out to find the source of source of this error. It helped to separate the model predictions and writing to a submission file into separate code snippets. To debug I did this: \n 1. Initialized the FVC values to 0 and confidence to 250. (Cell block 1)\n 2. Made the model predictions. (Cell block 2)\n 3. Saved them to submission file. (Cell block 3)\n\nTo my surprise, the submission scored `-11.1059` which is exact score one would get if the model predictions failed. \n\n**The Fix:** Removing a multi-line comment helped me fix the problematic code block. Multiline comments inside code blocks seemed to throw some exception during the notebook execution on the private test run (Works fine during commit though).\n\nSimilar issues for other people were resolved by:\n- Removing all files in the working directory, but for the submission.csv file.\n- Removing empty code cells and markdown cells.\n- Reading certain .dcm files: Try putting them inside a try-except block, make submission in seperate code block and see how it scores. Does it score near to your CV or does it score abnormally?\n\nHope this helps :)"
  },
  "source": "meta"
}