{ "id": 243559, "title": "Submussion Scoring Error", "url": "/competitions/siim-covid19-detection/discussion/243559", "author_name": "Nguyễn Tuấn Hiệp", "post_date": "2021-06-03T06:57:00.089000", "votes": 3, "comment_count": 10, "views": 0, "content": "

I have a problem with getting score when submit.
\nStatus is: Submussion Scoring Error and Public Score is error too.

", "messages": [ { "id": 1333940, "postDate": "2021-06-03T06:57:00.090Z", "content": "

I have a problem with getting score when submit.
\nStatus is: Submussion Scoring Error and Public Score is error too.

", "rawMarkdown": "I have a problem with getting score when submit.\nStatus is: Submussion Scoring Error and Public Score is error too.", "votes": 3 }, { "id": 1334075, "postDate": "2021-06-03T09:10:55.627Z", "content": "

yeah, me too. But I unerstand that obscuring the trace error helps to prevent cheating.

\n

It would be really helpful to have a way to \"Test\" your submission file against fake data, in this way there wouldn't be a risk of chating and you can see what's expected

", "rawMarkdown": "yeah, me too. But I unerstand that obscuring the trace error helps to prevent cheating. \n\nIt would be really helpful to have a way to \"Test\" your submission file against fake data, in this way there wouldn't be a risk of chating and you can see what's expected", "votes": 1, "replies": [ { "id": 1349003, "postDate": "2021-06-14T12:56:57.820Z", "content": "

try my new comment, may be it can help you.

", "rawMarkdown": "try my new comment, may be it can help you." } ] }, { "id": 1350474, "postDate": "2021-06-15T13:18:56.837Z", "content": "

I have the same problems trying to submit a locally generated submission file using a very simple notebook. So far the results are as follows.

\n
    \n
  1. Read the my generated submission csv to pandas and save it as submission.csv - \"Submission scoring error\"
    2. \n
  2. Read the sample submission and use pandas merge to merge with my submission on id - \"Submission file not found\"
    3. \n
  3. Iterate sequentially through all the records change the PredictionString value - \"Submission file not found\"
    4. \n
  4. Replace the PredictionString column from sample submission with the column from my submission - submissions gets very low score, lower even than the baseline.
    5. \n
\n

In all four cases the submission.csv files when downloaded from the corresponding notebook are identical (using linux diff to check)

", "rawMarkdown": "I have the same problems trying to submit a locally generated submission file using a very simple notebook. So far the results are as follows.\n\n1. Read the my generated submission csv to pandas and save it as submission.csv - \"Submission scoring error\"\n2. Read the sample submission and use pandas merge to merge with my submission on id - \"Submission file not found\"\n3. Iterate sequentially through all the records change the PredictionString value - \"Submission file not found\"\n4. Replace the PredictionString column from sample submission with the column from my submission - submissions gets very low score, lower even than the baseline.\n\nIn all four cases the submission.csv files when downloaded from the corresponding notebook are identical (using linux diff to check)", "replies": [ { "id": 1350525, "postDate": "2021-06-15T14:04:06.810Z", "content": "

When you click submit, the test dataset is replaced by a hidden test dataset. Your notebook should generate a submission file on that hidden test dataset when submitted. You can not generate a submission file locally as the hidden test dataset is not given. You can see some public inference kernels for example.

\n

You can look at https://www.kaggle.com/code-competition-debugging for details regarding the errors that you have encountered.

", "rawMarkdown": "When you click submit, the test dataset is replaced by a hidden test dataset. Your notebook should generate a submission file on that hidden test dataset when submitted. You can not generate a submission file locally as the hidden test dataset is not given. You can see some public inference kernels for example.\n\nYou can look at https://www.kaggle.com/code-competition-debugging for details regarding the errors that you have encountered. \n" } ] }, { "id": 1349000, "postDate": "2021-06-14T12:55:02.537Z", "content": "

I uploaded test folder converted from my pc to make execute fast and get score error in the past.

\n

Now, I scored. I read file sample_submission to get id and predict these id in test images folder which I converted directly from test folder dicom.

\n

I think it should help.

", "rawMarkdown": "I uploaded test folder converted from my pc to make execute fast and get score error in the past. \n\nNow, I scored. I read file sample_submission to get id and predict these id in test images folder which I converted directly from test folder dicom. \n\nI think it should help.", "replies": [ { "id": 1349002, "postDate": "2021-06-14T12:56:30.373Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 1334344, "postDate": "2021-06-03T12:40:30.200Z", "content": "

I currently have \"Submission CSV Not Found\" however I see in my kernel there is a submission.csv file in the output…

", "rawMarkdown": "I currently have \"Submission CSV Not Found\" however I see in my kernel there is a submission.csv file in the output...", "replies": [ { "id": 1348920, "postDate": "2021-06-14T11:46:13.187Z", "content": "

I am not sure if I was too late. But I have been with this problem for a long time. My fix was to ensure that there is \"submission.csv\" in \"/kaggle/working/\". But only that file alone, I ended up delete other intermediate files that I created. I still have problem submitting the result though (scoring error), so take this with the grain of salts.

", "rawMarkdown": "I am not sure if I was too late. But I have been with this problem for a long time. My fix was to ensure that there is \"submission.csv\" in \"/kaggle/working/\". But only that file alone, I ended up delete other intermediate files that I created. I still have problem submitting the result though (scoring error), so take this with the grain of salts." }, { "id": 1349001, "postDate": "2021-06-14T12:55:44.660Z", "content": "

try my new comment, may be it can help you.

", "rawMarkdown": "try my new comment, may be it can help you." }, { "id": 1349006, "postDate": "2021-06-14T12:57:55.790Z", "content": "

@Jacek Poplawski do as Theerit, ensure that there is only \"submission.csv\" in \"/kaggle/working/\"

", "rawMarkdown": "@Jacek Poplawski do as Theerit, ensure that there is only \"submission.csv\" in \"/kaggle/working/\"", "votes": 1 } ] } ], "comments": [ { "id": 1334075, "author_name": "Matias Thayer", "author_url": "", "post_date": "2021-06-03T09:10:55.627000", "content": "

yeah, me too. But I unerstand that obscuring the trace error helps to prevent cheating.

\n

It would be really helpful to have a way to \"Test\" your submission file against fake data, in this way there wouldn't be a risk of chating and you can see what's expected

", "votes": 1, "replies": [ { "id": 1349003, "author_name": "Nguyễn Tuấn Hiệp", "author_url": "", "post_date": "2021-06-14T12:56:57.820000", "content": "

try my new comment, may be it can help you.

", "votes": 0, "replies": [] } ] }, { "id": 1350474, "author_name": "Dmitriy Umerenkov", "author_url": "", "post_date": "2021-06-15T13:18:56.837000", "content": "

I have the same problems trying to submit a locally generated submission file using a very simple notebook. So far the results are as follows.

\n
    \n
  1. Read the my generated submission csv to pandas and save it as submission.csv - \"Submission scoring error\"
    2. \n
  2. Read the sample submission and use pandas merge to merge with my submission on id - \"Submission file not found\"
    3. \n
  3. Iterate sequentially through all the records change the PredictionString value - \"Submission file not found\"
    4. \n
  4. Replace the PredictionString column from sample submission with the column from my submission - submissions gets very low score, lower even than the baseline.
    5. \n
\n

In all four cases the submission.csv files when downloaded from the corresponding notebook are identical (using linux diff to check)

", "votes": 0, "replies": [ { "id": 1350525, "author_name": "Mohammad Zunaed", "author_url": "", "post_date": "2021-06-15T14:04:06.810000", "content": "

When you click submit, the test dataset is replaced by a hidden test dataset. Your notebook should generate a submission file on that hidden test dataset when submitted. You can not generate a submission file locally as the hidden test dataset is not given. You can see some public inference kernels for example.

\n

You can look at https://www.kaggle.com/code-competition-debugging for details regarding the errors that you have encountered.

", "votes": 0, "replies": [] } ] }, { "id": 1349000, "author_name": "Nguyễn Tuấn Hiệp", "author_url": "", "post_date": "2021-06-14T12:55:02.537000", "content": "

I uploaded test folder converted from my pc to make execute fast and get score error in the past.

\n

Now, I scored. I read file sample_submission to get id and predict these id in test images folder which I converted directly from test folder dicom.

\n

I think it should help.

", "votes": 0, "replies": [ { "id": 1349002, "author_name": "", "author_url": "", "post_date": "2021-06-14T12:56:30.373000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 1334344, "author_name": "Jacek Poplawski", "author_url": "", "post_date": "2021-06-03T12:40:30.200000", "content": "

I currently have \"Submission CSV Not Found\" however I see in my kernel there is a submission.csv file in the output…

", "votes": 0, "replies": [ { "id": 1348920, "author_name": "Theerit", "author_url": "", "post_date": "2021-06-14T11:46:13.187000", "content": "

I am not sure if I was too late. But I have been with this problem for a long time. My fix was to ensure that there is \"submission.csv\" in \"/kaggle/working/\". But only that file alone, I ended up delete other intermediate files that I created. I still have problem submitting the result though (scoring error), so take this with the grain of salts.

", "votes": 0, "replies": [] }, { "id": 1349001, "author_name": "Nguyễn Tuấn Hiệp", "author_url": "", "post_date": "2021-06-14T12:55:44.660000", "content": "

try my new comment, may be it can help you.

", "votes": 0, "replies": [] }, { "id": 1349006, "author_name": "Nguyễn Tuấn Hiệp", "author_url": "", "post_date": "2021-06-14T12:57:55.790000", "content": "

@Jacek Poplawski do as Theerit, ensure that there is only \"submission.csv\" in \"/kaggle/working/\"

", "votes": 1, "replies": [] } ] } ], "raw_markdown_by_id": { "1333940": "I have a problem with getting score when submit.\nStatus is: Submussion Scoring Error and Public Score is error too.", "1334075": "yeah, me too. But I unerstand that obscuring the trace error helps to prevent cheating. \n\nIt would be really helpful to have a way to \"Test\" your submission file against fake data, in this way there wouldn't be a risk of chating and you can see what's expected", "1350474": "I have the same problems trying to submit a locally generated submission file using a very simple notebook. So far the results are as follows.\n\n1. Read the my generated submission csv to pandas and save it as submission.csv - \"Submission scoring error\"\n2. Read the sample submission and use pandas merge to merge with my submission on id - \"Submission file not found\"\n3. Iterate sequentially through all the records change the PredictionString value - \"Submission file not found\"\n4. Replace the PredictionString column from sample submission with the column from my submission - submissions gets very low score, lower even than the baseline.\n\nIn all four cases the submission.csv files when downloaded from the corresponding notebook are identical (using linux diff to check)", "1349000": "I uploaded test folder converted from my pc to make execute fast and get score error in the past. \n\nNow, I scored. I read file sample_submission to get id and predict these id in test images folder which I converted directly from test folder dicom. \n\nI think it should help.", "1334344": "I currently have \"Submission CSV Not Found\" however I see in my kernel there is a submission.csv file in the output..." } }