Is the correct way to submit:
\n1) create a notebook and run i it on the pulblic test dataset to create a submissions.csv file,
\n2) then under submissions submit the same notebook to run it again?
The whole processes takes several hours (I set this off last night and found notebook hadnt finished as kaggle paused it with message \"are you still there, continue running\" - is there a way to turn that off?), just want to make sure this is the quickest way to submit (eg alternatively, if test length>x run whole notebook else run quick dummy code to produce dummy submission.csv)?.
", "messages": [ { "id": 1049296, "postDate": "2020-10-14T10:02:08.727Z", "content": "it's been eight hours still my public score not awarded still running🤕🤕
", "rawMarkdown": "it's been eight hours still my public score not awarded still running🤕🤕", "votes": 1 }, { "id": 1047366, "postDate": "2020-10-12T14:15:56.257Z", "content": "Correct. When submitting the notebook, public test set is replaced by a private dataset and your code is running simultaneously on private and public test images. This takes some time due to large sample sizes.
\nYou can speed up the notebook commit time by not predicting all images of the public test set during commit as suggested in this discussion topic. However, the submission time would not be affected by this trick.
", "rawMarkdown": "Correct. When submitting the notebook, public test set is replaced by a private dataset and your code is running simultaneously on private and public test images. This takes some time due to large sample sizes.\n\nYou can speed up the notebook commit time by not predicting all images of the public test set during commit as suggested in [this discussion topic](https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/discussion/186982). However, the submission time would not be affected by this trick.", "votes": 1, "replies": [ { "id": 1047741, "postDate": "2020-10-12T22:18:47.150Z", "content": "Thanks, the ability to run on a portion of public test set will speed things up.
", "rawMarkdown": "Thanks, the ability to run on a portion of public test set will speed things up." } ] }, { "id": 1047333, "postDate": "2020-10-12T13:42:20.097Z", "content": "Is the correct way to submit:
\n1) create a notebook and run i it on the pulblic test dataset to create a submissions.csv file,
\n2) then under submissions submit the same notebook to run it again?
The whole processes takes several hours (I set this off last night and found notebook hadnt finished as kaggle paused it with message \"are you still there, continue running\" - is there a way to turn that off?), just want to make sure this is the quickest way to submit (eg alternatively, if test length>x run whole notebook else run quick dummy code to produce dummy submission.csv)?.
", "rawMarkdown": "Is the correct way to submit: \n1) create a notebook and run i it on the pulblic test dataset to create a submissions.csv file, \n2) then under submissions submit the same notebook to run it again?\n\nThe whole processes takes several hours (I set this off last night and found notebook hadnt finished as kaggle paused it with message \"are you still there, continue running\" - is there a way to turn that off?), just want to make sure this is the quickest way to submit (eg alternatively, if test length>x run whole notebook else run quick dummy code to produce dummy submission.csv)?.", "votes": 1 }, { "id": 1047739, "postDate": "2020-10-12T22:14:48.003Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 1049296, "author_name": "Javagar M", "author_url": "", "post_date": "2020-10-14T10:02:08.727000", "content": "it's been eight hours still my public score not awarded still running🤕🤕
", "votes": 1, "replies": [] }, { "id": 1047366, "author_name": "Nikita Kozodoi", "author_url": "", "post_date": "2020-10-12T14:15:56.257000", "content": "Correct. When submitting the notebook, public test set is replaced by a private dataset and your code is running simultaneously on private and public test images. This takes some time due to large sample sizes.
\nYou can speed up the notebook commit time by not predicting all images of the public test set during commit as suggested in this discussion topic. However, the submission time would not be affected by this trick.
", "votes": 1, "replies": [ { "id": 1047741, "author_name": "Agedev", "author_url": "", "post_date": "2020-10-12T22:18:47.150000", "content": "Thanks, the ability to run on a portion of public test set will speed things up.
", "votes": 0, "replies": [] } ] }, { "id": 1047739, "author_name": "", "author_url": "", "post_date": "2020-10-12T22:14:48.003000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "1049296": "it's been eight hours still my public score not awarded still running🤕🤕", "1047366": "Correct. When submitting the notebook, public test set is replaced by a private dataset and your code is running simultaneously on private and public test images. This takes some time due to large sample sizes.\n\nYou can speed up the notebook commit time by not predicting all images of the public test set during commit as suggested in [this discussion topic](https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/discussion/186982). However, the submission time would not be affected by this trick.", "1047333": "Is the correct way to submit: \n1) create a notebook and run i it on the pulblic test dataset to create a submissions.csv file, \n2) then under submissions submit the same notebook to run it again?\n\nThe whole processes takes several hours (I set this off last night and found notebook hadnt finished as kaggle paused it with message \"are you still there, continue running\" - is there a way to turn that off?), just want to make sure this is the quickest way to submit (eg alternatively, if test length>x run whole notebook else run quick dummy code to produce dummy submission.csv)?.", "1047739": "" } }