Is there any way to know the reason a submission has thrown an error when no error is available?\nBesides what is the problem when the submission error is \"csv file not found\" and the csv file can be downloaded normally?\nAny help is appreciated.
", "messages": [ { "id": "954564", "postDate": "08/01/2020 20:58:48", "content": "Is there any way to know the reason a submission has thrown an error when no error is available?\nBesides what is the problem when the submission error is \"csv file not found\" and the csv file can be downloaded normally?\nAny help is appreciated.
", "rawMarkdown": "Is there any way to know the reason a submission has thrown an error when no error is available?\nBesides what is the problem when the submission error is \"csv file not found\" and the csv file can be downloaded normally?\nAny help is appreciated.", "votes": null }, { "id": "955112", "postDate": "08/02/2020 11:01:50", "content": "you have to name submission file as \"submission.csv\"
", "rawMarkdown": "you have to name submission file as \"submission.csv\"", "votes": null }, { "id": "955254", "postDate": "08/02/2020 13:46:02", "content": "Thanks for your answer.The file name is correct.If it were not,\nnotebook processing would have trapped this error, along with session internet connection, from the very begining.
", "rawMarkdown": "Thanks for your answer.The file name is correct.If it were not,\nnotebook processing would have trapped this error, along with session internet connection, from the very begining.", "votes": null }, { "id": "955419", "postDate": "08/02/2020 16:11:44", "content": "https://www.kaggle.com/c/global-wheat-detection/discussion/149809\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149871
\n\nThese are threads from other competitions for submission csv not found error. Hopefully they can provide information on what checks to be done for your case.
\n\nI would also advise to have (if not already done) a \"try except\" block where ever possible while generating the submission file.
", "rawMarkdown": "https://www.kaggle.com/c/global-wheat-detection/discussion/149809\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149871\n\nThese are threads from other competitions for submission csv not found error. Hopefully they can provide information on what checks to be done for your case.\n\nI would also advise to have (if not already done) a \"try except\" block where ever possible while generating the submission file.", "votes": null }, { "id": "955650", "postDate": "08/02/2020 19:22:18", "content": "Thanks for the info
", "rawMarkdown": "Thanks for the info", "votes": null }, { "id": "956093", "postDate": "08/03/2020 08:29:55", "content": "Have you found any solution? I am getting submission error for all my submissions. I generate the \"submission.csv\" file, disabled internet, and checked the output that my columns are named correctly. Also removed the index column, which was not supposed to be in the submission.csv file. Also tried to format the FVC as integer only but other kernels seem to submit fine with floats there..
\n\nYet I still only get the \"Submission Scoring Error\". Also the little (i) info button next to \"Error\" column just says \"no additional data provided for this error\".
\n\nThere must be some log somewhere?
", "rawMarkdown": "Have you found any solution? I am getting submission error for all my submissions. I generate the \"submission.csv\" file, disabled internet, and checked the output that my columns are named correctly. Also removed the index column, which was not supposed to be in the submission.csv file. Also tried to format the FVC as integer only but other kernels seem to submit fine with floats there..\n\nYet I still only get the \"Submission Scoring Error\". Also the little (i) info button next to \"Error\" column just says \"no additional data provided for this error\".\n\nThere must be some log somewhere?", "votes": null }, { "id": "956114", "postDate": "08/03/2020 08:49:24", "content": "@donkeys Unfortunately there is no log. I have faced this problem in another competition - https://www.kaggle.com/c/global-wheat-detection/discussion/167624
\n\nThere were two things causing the problem for me:
\n\nOne other thing I tried and was not sure if it made any difference was - I has my own datasets for offline packages and pretrained weights which were private before and I made them to public.
", "rawMarkdown": "donkeys Unfortunately there is no log. I have faced this problem in another competition - https://www.kaggle.com/c/global-wheat-detection/discussion/167624\n\nThere were two things causing the problem for me:\n\n - I was doing \"Quick Save\" most of the times while saving the versions while the instructions say that we have to do \"Save and Run all\"\n - I was adding RGB values of all test images into a list to visualise the object detection using matplotlib, as the public test images were only 10, I never got an error when ran on 10 images but there were 1000s of images during submission it failed as CUDA / CPU was not sufficient. **So its worth a while to examine all variables being used and ensure they do not cause memory crashes for large data during submission.**\n\nOne other thing I tried and was not sure if it made any difference was - I has my own datasets for offline packages and pretrained weights which were private before and I made them to public.", "votes": null }, { "id": "956132", "postDate": "08/03/2020 09:14:05", "content": "Unfortunately no @averagemn.I rerun my last acceptable solution and it passed, in order to exclude any local issues.I checked the shapes and the formats with no result.\nI suspect numerical issues in the file may cause trouble .\nThe oddiest think is that sometimes I get \"No csv found\" and the file is there.\nReading the discussions I see a lot of people have the same problem. \nMaybe some kaggle admin can sheed some light...
", "rawMarkdown": "Unfortunately no @averagemn.I rerun my last acceptable solution and it passed, in order to exclude any local issues.I checked the shapes and the formats with no result.\nI suspect numerical issues in the file may cause trouble .\nThe oddiest think is that sometimes I get \"No csv found\" and the file is there.\nReading the discussions I see a lot of people have the same problem. \nMaybe some kaggle admin can sheed some light...", "votes": null }, { "id": "956446", "postDate": "08/03/2020 14:05:33", "content": "My progress report:\n1..The clip in confidence is active in submissions. If a value is less than 70 you get thrown out.\n2..There must be a check in ( FVC - confidence) .If it is less than 2000 you get thrown out too.\nPlease check the above in your cases, at least they apply to me .
", "rawMarkdown": "My progress report:\n1..The clip in confidence is active in submissions. If a value is less than 70 you get thrown out.\n2..There must be a check in ( FVC - confidence) .If it is less than 2000 you get thrown out too.\nPlease check the above in your cases, at least they apply to me .", "votes": null }, { "id": "956474", "postDate": "08/03/2020 14:31:29", "content": "Thanks. It's probably the confidence values, since I was experimenting and have some work to finetune those still. Wish these limits were clearly documented and there was some actual logs to see what is happening..
\n\nI will try to modify my outputs and see how it works. Just takes while since it runs the entire kernel on submission.
", "rawMarkdown": "Thanks. It's probably the confidence values, since I was experimenting and have some work to finetune those still. Wish these limits were clearly documented and there was some actual logs to see what is happening..\n\nI will try to modify my outputs and see how it works. Just takes while since it runs the entire kernel on submission.", "votes": null }, { "id": "958363", "postDate": "08/05/2020 00:24:25", "content": "There are a few threads about this same issue so I'll copy-paste my answer here:
\n\n\n", "rawMarkdown": "There are a few threads about this same issue so I'll copy-paste my answer here:\n\n> For a while I also had issues submitting. The thing is, if while running on the private data set your code produces an error in the cell that outputs your CSV, the execution stops and no CSV is created.\n\n> The way I found to debug this sort of issue was by running the code that produces the predictions and submission CSV on the *training* set (filtered to keep only one record per patient), which presents a greater range of potential issues than the 5-patient public test set. This allowed me to access a normal error log and zero in on what exactly was causing a problem.", "votes": null } ], "comments": [ { "id": 955112, "author_name": "suchith0312", "author_url": "", "post_date": "08/02/2020 11:01:50", "content": "For a while I also had issues submitting. The thing is, if while running on the private data set your code produces an error in the cell that outputs your CSV, the execution stops and no CSV is created.
\n \nThe way I found to debug this sort of issue was by running the code that produces the predictions and submission CSV on the training set (filtered to keep only one record per patient), which presents a greater range of potential issues than the 5-patient public test set. This allowed me to access a normal error log and zero in on what exactly was causing a problem.
\n
you have to name submission file as \"submission.csv\"
", "votes": null, "replies": [ { "id": 955254, "author_name": "georgem", "author_url": "", "post_date": "08/02/2020 13:46:02", "content": "Thanks for your answer.The file name is correct.If it were not,\nnotebook processing would have trapped this error, along with session internet connection, from the very begining.
", "votes": null, "replies": [] } ] }, { "id": 955419, "author_name": "havinath", "author_url": "", "post_date": "08/02/2020 16:11:44", "content": "https://www.kaggle.com/c/global-wheat-detection/discussion/149809\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149871
\n\nThese are threads from other competitions for submission csv not found error. Hopefully they can provide information on what checks to be done for your case.
\n\nI would also advise to have (if not already done) a \"try except\" block where ever possible while generating the submission file.
", "votes": null, "replies": [ { "id": 955650, "author_name": "georgem", "author_url": "", "post_date": "08/02/2020 19:22:18", "content": "Thanks for the info
", "votes": null, "replies": [] } ] }, { "id": 956093, "author_name": "donkeys", "author_url": "", "post_date": "08/03/2020 08:29:55", "content": "Have you found any solution? I am getting submission error for all my submissions. I generate the \"submission.csv\" file, disabled internet, and checked the output that my columns are named correctly. Also removed the index column, which was not supposed to be in the submission.csv file. Also tried to format the FVC as integer only but other kernels seem to submit fine with floats there..
\n\nYet I still only get the \"Submission Scoring Error\". Also the little (i) info button next to \"Error\" column just says \"no additional data provided for this error\".
\n\nThere must be some log somewhere?
", "votes": null, "replies": [ { "id": 956114, "author_name": "havinath", "author_url": "", "post_date": "08/03/2020 08:49:24", "content": "@donkeys Unfortunately there is no log. I have faced this problem in another competition - https://www.kaggle.com/c/global-wheat-detection/discussion/167624
\n\nThere were two things causing the problem for me:
\n\nOne other thing I tried and was not sure if it made any difference was - I has my own datasets for offline packages and pretrained weights which were private before and I made them to public.
", "votes": null, "replies": [] }, { "id": 956132, "author_name": "georgem", "author_url": "", "post_date": "08/03/2020 09:14:05", "content": "Unfortunately no @averagemn.I rerun my last acceptable solution and it passed, in order to exclude any local issues.I checked the shapes and the formats with no result.\nI suspect numerical issues in the file may cause trouble .\nThe oddiest think is that sometimes I get \"No csv found\" and the file is there.\nReading the discussions I see a lot of people have the same problem. \nMaybe some kaggle admin can sheed some light...
", "votes": null, "replies": [] }, { "id": 956446, "author_name": "georgem", "author_url": "", "post_date": "08/03/2020 14:05:33", "content": "My progress report:\n1..The clip in confidence is active in submissions. If a value is less than 70 you get thrown out.\n2..There must be a check in ( FVC - confidence) .If it is less than 2000 you get thrown out too.\nPlease check the above in your cases, at least they apply to me .
", "votes": null, "replies": [] }, { "id": 956474, "author_name": "donkeys", "author_url": "", "post_date": "08/03/2020 14:31:29", "content": "Thanks. It's probably the confidence values, since I was experimenting and have some work to finetune those still. Wish these limits were clearly documented and there was some actual logs to see what is happening..
\n\nI will try to modify my outputs and see how it works. Just takes while since it runs the entire kernel on submission.
", "votes": null, "replies": [] }, { "id": 958363, "author_name": "archimedus", "author_url": "", "post_date": "08/05/2020 00:24:25", "content": "There are a few threads about this same issue so I'll copy-paste my answer here:
\n\n\n", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "954564": "Is there any way to know the reason a submission has thrown an error when no error is available?\nBesides what is the problem when the submission error is \"csv file not found\" and the csv file can be downloaded normally?\nAny help is appreciated.", "955112": "you have to name submission file as \"submission.csv\"", "955254": "Thanks for your answer.The file name is correct.If it were not,\nnotebook processing would have trapped this error, along with session internet connection, from the very begining.", "955419": "https://www.kaggle.com/c/global-wheat-detection/discussion/149809\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149871\n\nThese are threads from other competitions for submission csv not found error. Hopefully they can provide information on what checks to be done for your case.\n\nI would also advise to have (if not already done) a \"try except\" block where ever possible while generating the submission file.", "955650": "Thanks for the info", "956093": "Have you found any solution? I am getting submission error for all my submissions. I generate the \"submission.csv\" file, disabled internet, and checked the output that my columns are named correctly. Also removed the index column, which was not supposed to be in the submission.csv file. Also tried to format the FVC as integer only but other kernels seem to submit fine with floats there..\n\nYet I still only get the \"Submission Scoring Error\". Also the little (i) info button next to \"Error\" column just says \"no additional data provided for this error\".\n\nThere must be some log somewhere?", "956114": "donkeys Unfortunately there is no log. I have faced this problem in another competition - https://www.kaggle.com/c/global-wheat-detection/discussion/167624\n\nThere were two things causing the problem for me:\n\n - I was doing \"Quick Save\" most of the times while saving the versions while the instructions say that we have to do \"Save and Run all\"\n - I was adding RGB values of all test images into a list to visualise the object detection using matplotlib, as the public test images were only 10, I never got an error when ran on 10 images but there were 1000s of images during submission it failed as CUDA / CPU was not sufficient. **So its worth a while to examine all variables being used and ensure they do not cause memory crashes for large data during submission.**\n\nOne other thing I tried and was not sure if it made any difference was - I has my own datasets for offline packages and pretrained weights which were private before and I made them to public.", "956132": "Unfortunately no @averagemn.I rerun my last acceptable solution and it passed, in order to exclude any local issues.I checked the shapes and the formats with no result.\nI suspect numerical issues in the file may cause trouble .\nThe oddiest think is that sometimes I get \"No csv found\" and the file is there.\nReading the discussions I see a lot of people have the same problem. \nMaybe some kaggle admin can sheed some light...", "956446": "My progress report:\n1..The clip in confidence is active in submissions. If a value is less than 70 you get thrown out.\n2..There must be a check in ( FVC - confidence) .If it is less than 2000 you get thrown out too.\nPlease check the above in your cases, at least they apply to me .", "956474": "Thanks. It's probably the confidence values, since I was experimenting and have some work to finetune those still. Wish these limits were clearly documented and there was some actual logs to see what is happening..\n\nI will try to modify my outputs and see how it works. Just takes while since it runs the entire kernel on submission.", "958363": "There are a few threads about this same issue so I'll copy-paste my answer here:\n\n> For a while I also had issues submitting. The thing is, if while running on the private data set your code produces an error in the cell that outputs your CSV, the execution stops and no CSV is created.\n\n> The way I found to debug this sort of issue was by running the code that produces the predictions and submission CSV on the *training* set (filtered to keep only one record per patient), which presents a greater range of potential issues than the 5-patient public test set. This allowed me to access a normal error log and zero in on what exactly was causing a problem." }, "source": "meta" }For a while I also had issues submitting. The thing is, if while running on the private data set your code produces an error in the cell that outputs your CSV, the execution stops and no CSV is created.
\n \nThe way I found to debug this sort of issue was by running the code that produces the predictions and submission CSV on the training set (filtered to keep only one record per patient), which presents a greater range of potential issues than the 5-patient public test set. This allowed me to access a normal error log and zero in on what exactly was causing a problem.
\n