![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1904745%2F0d5ba53e3f5fecc7b5a5b5509669984c%2FSans%20titre.png?generation=1598827276767654&alt=media\")
Hi i a using output result from my others commited notebooks :
\nhttps://www.kaggle.com/servietsky/merge-dicom-csv-and-ct-scans
\nhttps://www.kaggle.com/servietsky/ct-scans-to-dataframe
\nAnd i use it for submit my predictions, however i have Submission scoring error, although I have no errors in the local execution.
\nAs kaggle evaluates the private LB at commit time using new test data that we do not have access to, so the preprocess I did on my old notebooks does not apply to new test dataset for private LB.
\nCould this be the cause of the submission error?
", "messages": [ { "id": "989270", "postDate": "08/28/2020 17:13:48", "content": "Hi i a using output result from my others commited notebooks :
\nhttps://www.kaggle.com/servietsky/merge-dicom-csv-and-ct-scans
\nhttps://www.kaggle.com/servietsky/ct-scans-to-dataframe
\nAnd i use it for submit my predictions, however i have Submission scoring error, although I have no errors in the local execution.
\nAs kaggle evaluates the private LB at commit time using new test data that we do not have access to, so the preprocess I did on my old notebooks does not apply to new test dataset for private LB.
\nCould this be the cause of the submission error?
", "rawMarkdown": "Hi i a using output result from my others commited notebooks :\n\nhttps://www.kaggle.com/servietsky/merge-dicom-csv-and-ct-scans\n\nhttps://www.kaggle.com/servietsky/ct-scans-to-dataframe\n\nAnd i use it for submit my predictions, however i have Submission scoring error, although I have no errors in the local execution.\n\nAs kaggle evaluates the private LB at commit time using new test data that we do not have access to, so the preprocess I did on my old notebooks does not apply to new test dataset for private LB.\n\nCould this be the cause of the submission error?", "votes": null }, { "id": "990975", "postDate": "08/30/2020 02:44:10", "content": "I've found this
\n\n", "rawMarkdown": "I've found this\n\n> This is a synchronous rerun code competition. The provided test set is a small representative set of files (copied from the training set) to demonstrate the format of the private test set. When you submit your notebook, Kaggle will rerun your code on the test set, which contains unseen images.", "votes": null }, { "id": "991306", "postDate": "08/30/2020 09:57:52", "content": "This is a synchronous rerun code competition. The provided test set is a small representative set of files (copied from the training set) to demonstrate the format of the private test set. When you submit your notebook, Kaggle will rerun your code on the test set, which contains unseen images.
\n
Hum… so we have to execute all preprocess and prediction in same notboook ?
", "rawMarkdown": "Hum... so we have to execute all preprocess and prediction in same notboook ?", "votes": null }, { "id": "991460", "postDate": "08/30/2020 12:28:53", "content": "I'm not sure but try putting all your work together.
", "rawMarkdown": "I'm not sure but try putting all your work together.", "votes": null }, { "id": "991580", "postDate": "08/30/2020 13:59:12", "content": "Check your submission file because it's must be '.*csv' and you shouldn't use encoded targets
", "rawMarkdown": "Check your submission file because it's must be '.*csv' and you shouldn't use encoded targets", "votes": null }, { "id": "991612", "postDate": "08/30/2020 14:31:10", "content": "Correct. You have no access to the real output of your notebook. You cannot blend separate notebooks.
", "rawMarkdown": "Correct. You have no access to the real output of your notebook. You cannot blend separate notebooks.", "votes": null }, { "id": "992083", "postDate": "08/30/2020 22:36:04", "content": "haha u think im doing that x), bad faith dont under estimate
", "rawMarkdown": "haha u think im doing that x), bad faith dont under estimate", "votes": null }, { "id": "992085", "postDate": "08/30/2020 22:41:20", "content": "i put only test preprocess in same notebook, in local it took 30 minutes runing for preprocess, predict and create submition but now it is taking more then 4 hours runing for evaluation.
\n
i checked that, it is CSV file
", "rawMarkdown": "i checked that, it is CSV file", "votes": null } ], "comments": [ { "id": 990975, "author_name": "hiramcho", "author_url": "", "post_date": "08/30/2020 02:44:10", "content": "I've found this
\n\n", "votes": null, "replies": [ { "id": 991306, "author_name": "servietsky", "author_url": "", "post_date": "08/30/2020 09:57:52", "content": "This is a synchronous rerun code competition. The provided test set is a small representative set of files (copied from the training set) to demonstrate the format of the private test set. When you submit your notebook, Kaggle will rerun your code on the test set, which contains unseen images.
\n
Hum… so we have to execute all preprocess and prediction in same notboook ?
", "votes": null, "replies": [ { "id": 991612, "author_name": "richardepstein", "author_url": "", "post_date": "08/30/2020 14:31:10", "content": "Correct. You have no access to the real output of your notebook. You cannot blend separate notebooks.
", "votes": null, "replies": [] }, { "id": 992083, "author_name": "servietsky", "author_url": "", "post_date": "08/30/2020 22:36:04", "content": "haha u think im doing that x), bad faith dont under estimate
", "votes": null, "replies": [] } ] }, { "id": 991460, "author_name": "hiramcho", "author_url": "", "post_date": "08/30/2020 12:28:53", "content": "I'm not sure but try putting all your work together.
", "votes": null, "replies": [] }, { "id": 992085, "author_name": "servietsky", "author_url": "", "post_date": "08/30/2020 22:41:20", "content": "i put only test preprocess in same notebook, in local it took 30 minutes runing for preprocess, predict and create submition but now it is taking more then 4 hours runing for evaluation.
\n
Check your submission file because it's must be '.*csv' and you shouldn't use encoded targets
", "votes": null, "replies": [ { "id": 992086, "author_name": "servietsky", "author_url": "", "post_date": "08/30/2020 22:41:34", "content": "i checked that, it is CSV file
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "989270": "Hi i a using output result from my others commited notebooks :\n\nhttps://www.kaggle.com/servietsky/merge-dicom-csv-and-ct-scans\n\nhttps://www.kaggle.com/servietsky/ct-scans-to-dataframe\n\nAnd i use it for submit my predictions, however i have Submission scoring error, although I have no errors in the local execution.\n\nAs kaggle evaluates the private LB at commit time using new test data that we do not have access to, so the preprocess I did on my old notebooks does not apply to new test dataset for private LB.\n\nCould this be the cause of the submission error?", "990975": "I've found this\n\n> This is a synchronous rerun code competition. The provided test set is a small representative set of files (copied from the training set) to demonstrate the format of the private test set. When you submit your notebook, Kaggle will rerun your code on the test set, which contains unseen images.", "991306": "Hum... so we have to execute all preprocess and prediction in same notboook ?", "991460": "I'm not sure but try putting all your work together.", "991580": "Check your submission file because it's must be '.*csv' and you shouldn't use encoded targets", "991612": "Correct. You have no access to the real output of your notebook. You cannot blend separate notebooks.", "992083": "haha u think im doing that x), bad faith dont under estimate", "992085": "i put only test preprocess in same notebook, in local it took 30 minutes runing for preprocess, predict and create submition but now it is taking more then 4 hours runing for evaluation. \n\n", "992086": "i checked that, it is CSV file" }, "source": "meta" }