{ "id": 500262, "title": "How do you submit now that the test set is closed?", "url": "/competitions/home-credit-credit-risk-model-stability/discussion/500262", "author_name": "", "post_date": "2024-05-04T22:49:40.979436200Z", "votes": 1, "comment_count": 6, "views": 0, "content": "

This is my first competition so forgive me if I missed something, I really did try to look around on the code and discussion. If the test_base.csv is limited to 10 rows, how are we supposed to set up the notebook when we submit? Is the starter notebook not relevant anymore? Is there an example how we are supposed to submit now?

", "messages": [ { "id": "2793768", "postDate": "05/04/2024 22:49:40", "content": "

This is my first competition so forgive me if I missed something, I really did try to look around on the code and discussion. If the test_base.csv is limited to 10 rows, how are we supposed to set up the notebook when we submit? Is the starter notebook not relevant anymore? Is there an example how we are supposed to submit now?

", "rawMarkdown": "This is my first competition so forgive me if I missed something, I really did try to look around on the code and discussion. If the test_base.csv is limited to 10 rows, how are we supposed to set up the notebook when we submit? Is the starter notebook not relevant anymore? Is there an example how we are supposed to submit now?", "votes": null }, { "id": "2793812", "postDate": "05/05/2024 01:05:14", "content": "

After notebook submission it will be rerun to generate predictions and test_base.csv (plus other test files) will be replaced with complete versions. No special setup required

", "rawMarkdown": "After notebook submission it will be rerun to generate predictions and test_base.csv (plus other test files) will be replaced with complete versions. No special setup required", "votes": null }, { "id": "2793818", "postDate": "05/05/2024 01:24:25", "content": "

Got it thanks! Now I have a new problem, the submission goes OOM AFTER the submission is successfully run. Does that mean it ran fine with 10 row test dataset, and now with the 30% of the test set, it goes OOM?

", "rawMarkdown": "Got it thanks! Now I have a new problem, the submission goes OOM AFTER the submission is successfully run. Does that mean it ran fine with 10 row test dataset, and now with the 30% of the test set, it goes OOM?", "votes": null }, { "id": "2793966", "postDate": "05/05/2024 05:07:10", "content": "

@pustoi please check your code for bugs and memory issues- this indicates that your notebook used more than available memory. This error notebook will not ne used for the final evaluation.

", "rawMarkdown": "pustoi please check your code for bugs and memory issues- this indicates that your notebook used more than available memory. This error notebook will not ne used for the final evaluation.", "votes": null }, { "id": "2794440", "postDate": "05/05/2024 09:42:31", "content": "

It runs on full test set, but LB is based on just 30% of it. So delete your train set variable from memory before processing the test set to have memory for it

", "rawMarkdown": "It runs on full test set, but LB is based on just 30% of it. So delete your train set variable from memory before processing the test set to have memory for it", "votes": null }, { "id": "2799244", "postDate": "05/07/2024 16:50:27", "content": "

how does test_base.csv will be generated ? by using exiting data in training data case id's ? if new case_ids how does it will create feature engineering !! if we are not automated every thing , lot of confusion for me on test_base !!

", "rawMarkdown": "how does test_base.csv will be generated ? by using exiting data in training data case id's ? if new case_ids how does it will create feature engineering !! if we are not automated every thing , lot of confusion for me on test_base !!", "votes": null }, { "id": "2801084", "postDate": "05/08/2024 13:24:36", "content": "

Test set has new case ids.
\nFeature engineering has nothing to do with them as it's based on features (columns) and not case ids (rows)

", "rawMarkdown": "Test set has new case ids. \nFeature engineering has nothing to do with them as it's based on features (columns) and not case ids (rows)", "votes": null } ], "comments": [ { "id": 2793812, "author_name": "vitalykudelya", "author_url": "", "post_date": "05/05/2024 01:05:14", "content": "

After notebook submission it will be rerun to generate predictions and test_base.csv (plus other test files) will be replaced with complete versions. No special setup required

", "votes": null, "replies": [ { "id": 2793818, "author_name": "pustoi", "author_url": "", "post_date": "05/05/2024 01:24:25", "content": "

Got it thanks! Now I have a new problem, the submission goes OOM AFTER the submission is successfully run. Does that mean it ran fine with 10 row test dataset, and now with the 30% of the test set, it goes OOM?

", "votes": null, "replies": [ { "id": 2793966, "author_name": "ravi20076", "author_url": "", "post_date": "05/05/2024 05:07:10", "content": "

@pustoi please check your code for bugs and memory issues- this indicates that your notebook used more than available memory. This error notebook will not ne used for the final evaluation.

", "votes": null, "replies": [] }, { "id": 2794440, "author_name": "eu1234", "author_url": "", "post_date": "05/05/2024 09:42:31", "content": "

It runs on full test set, but LB is based on just 30% of it. So delete your train set variable from memory before processing the test set to have memory for it

", "votes": null, "replies": [] } ] }, { "id": 2799244, "author_name": "kunduruanil", "author_url": "", "post_date": "05/07/2024 16:50:27", "content": "

how does test_base.csv will be generated ? by using exiting data in training data case id's ? if new case_ids how does it will create feature engineering !! if we are not automated every thing , lot of confusion for me on test_base !!

", "votes": null, "replies": [ { "id": 2801084, "author_name": "eu1234", "author_url": "", "post_date": "05/08/2024 13:24:36", "content": "

Test set has new case ids.
\nFeature engineering has nothing to do with them as it's based on features (columns) and not case ids (rows)

", "votes": null, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2793768": "This is my first competition so forgive me if I missed something, I really did try to look around on the code and discussion. If the test_base.csv is limited to 10 rows, how are we supposed to set up the notebook when we submit? Is the starter notebook not relevant anymore? Is there an example how we are supposed to submit now?", "2793812": "After notebook submission it will be rerun to generate predictions and test_base.csv (plus other test files) will be replaced with complete versions. No special setup required", "2793818": "Got it thanks! Now I have a new problem, the submission goes OOM AFTER the submission is successfully run. Does that mean it ran fine with 10 row test dataset, and now with the 30% of the test set, it goes OOM?", "2793966": "pustoi please check your code for bugs and memory issues- this indicates that your notebook used more than available memory. This error notebook will not ne used for the final evaluation.", "2794440": "It runs on full test set, but LB is based on just 30% of it. So delete your train set variable from memory before processing the test set to have memory for it", "2799244": "how does test_base.csv will be generated ? by using exiting data in training data case id's ? if new case_ids how does it will create feature engineering !! if we are not automated every thing , lot of confusion for me on test_base !!", "2801084": "Test set has new case ids. \nFeature engineering has nothing to do with them as it's based on features (columns) and not case ids (rows)" }, "source": "meta" }