![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4629732%2Fd365ca703217a54c35d223b2999aa153%2Fnot_found.png?generation=1599199692869509&alt=media\")
How to debug a \"Submission CSV Not Found\" error
\n(when no further information is given)
\nI recently got this error when I tried to submit my results and Kaggle only gives the following \"hint\":
\n\n\"No additional details provided for this error\"
\n
Possible approach to find bugs like this:
\nIn his post https://www.kaggle.com/adaubas/a-solution-to-submissions-csv-not-found, @adaubas describes an unsual, but very creative solution to debug problems like this.
Problem symptoms
\nsubmission.csv like the given example submission)![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4629732%2F8ab61261ff578876f20414b87245d277%2Fimage.png?generation=1599198767789321&alt=media\")
Debug procedure (more in @adaubas 's post, s. above)
\ncreate_test_submission, that takes in varying submission values as an argument and submits static submission.csv files with these varying valuescreate_test_submission at different lines in the code and use it as a kind of breakpointHow to debug a \"Submission CSV Not Found\" error
\n(when no further information is given)
\nI recently got this error when I tried to submit my results and Kaggle only gives the following \"hint\":
\n\n\"No additional details provided for this error\"
\n
Possible approach to find bugs like this:
\nIn his post https://www.kaggle.com/adaubas/a-solution-to-submissions-csv-not-found, @adaubas describes an unsual, but very creative solution to debug problems like this.
Problem symptoms
\nsubmission.csv like the given example submission)
Debug procedure (more in @adaubas 's post, s. above)
\ncreate_test_submission, that takes in varying submission values as an argument and submits static submission.csv files with these varying valuescreate_test_submission at different lines in the code and use it as a kind of breakpointUpdate
\nI still have not found the error that causes my notebook to crash in \"submission mode\".
\nIn my case, it has something to do with the processing of the CT scans.
I assume that some of the scans in the hidden validation set that Kaggle plugs into my model cause the error. At the end of the day, it is really annoying that debugging is nearly impossible this way.
\nIf somebody has other ideas, please let me know!
", "rawMarkdown": "**Update**\nI still have not found the error that causes my notebook to crash in \"submission mode\".\nIn my case, it has something to do with the processing of the CT scans. \n\nI assume that some of the scans in the hidden validation set that Kaggle plugs into my model cause the error. At the end of the day, it is really annoying that debugging is nearly impossible this way.\n\nIf somebody has other ideas, please let me know!", "votes": null }, { "id": "1000740", "postDate": "09/06/2020 18:49:32", "content": "Always use try-except clauses if you're not sure, if any CT scan reading fails, you can populate that data with mean values (if features), or black masks (if using slices directly), it won't tell you where it failed but it will at least run the submission and give you a score.
", "rawMarkdown": "Always use try-except clauses if you're not sure, if any CT scan reading fails, you can populate that data with mean values (if features), or black masks (if using slices directly), it won't tell you where it failed but it will at least run the submission and give you a score.", "votes": null } ], "comments": [ { "id": 1000258, "author_name": "andihsg666", "author_url": "", "post_date": "09/06/2020 12:16:21", "content": "Update
\nI still have not found the error that causes my notebook to crash in \"submission mode\".
\nIn my case, it has something to do with the processing of the CT scans.
I assume that some of the scans in the hidden validation set that Kaggle plugs into my model cause the error. At the end of the day, it is really annoying that debugging is nearly impossible this way.
\nIf somebody has other ideas, please let me know!
", "votes": null, "replies": [ { "id": 1000740, "author_name": "furcifer", "author_url": "", "post_date": "09/06/2020 18:49:32", "content": "Always use try-except clauses if you're not sure, if any CT scan reading fails, you can populate that data with mean values (if features), or black masks (if using slices directly), it won't tell you where it failed but it will at least run the submission and give you a score.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "997613": "**How to debug a \"Submission CSV Not Found\" error** \n(when no further information is given)\nI recently got this error when I tried to submit my results and Kaggle only gives the following \"hint\":\n> \"No additional details provided for this error\"\n\n\n\n**Possible approach to find bugs like this:**\nIn his post https://www.kaggle.com/adaubas/a-solution-to-submissions-csv-not-found, @adaubas describes an unsual, but very creative solution to debug problems like this.\n\n**Problem symptoms**\n- You have already submitted a static (i.e., hard-coded `submission.csv` like the given example submission)\n- Notebook commit is successful\n- Notebook runs in editor without errors\n- Notebook produces output\n- Output file has the right file name and content, i.e. number of rows, values correct, etc.\n- Output file can be chosen for submission (s. screen shot below)\n\n\n\n**Debug procedure** (more in @adaubas 's post, s. above)\n- Implement a function `create_test_submission`, that takes in varying submission values as an argument and submits static `submission.csv` files with these **varying values**\n- Call `create_test_submission` at different lines in the code and use it as a kind of **breakpoint**\n- Your submitted values translate to your public leaderbord score, which will be lower or higher, depending on which submission file was created the last\n- From the displayed score in the leaderboard you can trace back the **last breakpoint** that created the last successful submission", "1000258": "**Update**\nI still have not found the error that causes my notebook to crash in \"submission mode\".\nIn my case, it has something to do with the processing of the CT scans. \n\nI assume that some of the scans in the hidden validation set that Kaggle plugs into my model cause the error. At the end of the day, it is really annoying that debugging is nearly impossible this way.\n\nIf somebody has other ideas, please let me know!", "1000740": "Always use try-except clauses if you're not sure, if any CT scan reading fails, you can populate that data with mean values (if features), or black masks (if using slices directly), it won't tell you where it failed but it will at least run the submission and give you a score." }, "source": "meta" }