{ "id": 191138, "title": "able to call make_env() multiple times", "url": "/competitions/riiid-test-answer-prediction/discussion/191138", "author_name": "", "post_date": "2020-10-14T21:02:26.507364600Z", "votes": 5, "comment_count": 5, "views": 0, "content": "

My notebook https://www.kaggle.com/yihdarshieh/able-to-call-make-env-multiple-times/ shows how to call make_env() multiple times.

\n

IMPORTANT This only works in interaction mode or commit mode. Submission will fail.

\n

This notebook shows you how to call make_env() multiple times in the same notebook session. It works also when we commit the notebook. However, the submission will fail (it will run for 9 hours and timeout).

\n

But while you build the input pipeline or models, this might make your work easier (without restarting the notebook).

\n

I hope this hack won't never work for submission! If someone smart enough (not me 😓) to do it, i hope he/she will contact the host.

", "messages": [ { "id": "1049864", "postDate": "10/14/2020 21:02:26", "content": "

My notebook https://www.kaggle.com/yihdarshieh/able-to-call-make-env-multiple-times/ shows how to call make_env() multiple times.

\n

IMPORTANT This only works in interaction mode or commit mode. Submission will fail.

\n

This notebook shows you how to call make_env() multiple times in the same notebook session. It works also when we commit the notebook. However, the submission will fail (it will run for 9 hours and timeout).

\n

But while you build the input pipeline or models, this might make your work easier (without restarting the notebook).

\n

I hope this hack won't never work for submission! If someone smart enough (not me 😓) to do it, i hope he/she will contact the host.

", "rawMarkdown": "My notebook [https://www.kaggle.com/yihdarshieh/able-to-call-make-env-multiple-times/](https://www.kaggle.com/yihdarshieh/able-to-call-make-env-multiple-times/) shows how to call `make_env()` multiple times.\n\n**IMPORTANT** This only works in interaction mode or commit mode. Submission will fail.\n\nThis notebook shows you how to call make_env() multiple times in the same notebook session. It works also when we commit the notebook. However, the submission will fail (it will run for 9 hours and timeout).\n\nBut while you build the input pipeline or models, this might make your work easier (without restarting the notebook).\n\nI hope this hack won't never work for submission! If someone smart enough (not me 😓) to do it, i hope he/she will contact the host.", "votes": null }, { "id": "1049894", "postDate": "10/14/2020 21:38:26", "content": "

I have a question for you. When your notebook failed after 9 hours, did you receive a submission scoring error or a different error?

", "rawMarkdown": "I have a question for you. When your notebook failed after 9 hours, did you receive a submission scoring error or a different error?", "votes": null }, { "id": "1049895", "postDate": "10/14/2020 21:39:38", "content": "

timeout error

", "rawMarkdown": "timeout error", "votes": null }, { "id": "1049897", "postDate": "10/14/2020 21:48:48", "content": "

Creative approach! Just a word of caution that testing ways to hack the API in a real submission may look sufficiently like a cheating attempt to us that we end up making a ban at the end of the competition. That's true regardless of whether or not the method actually ends up working.

\n

With that said, @yihdarshieh you don't need to worry about getting banned for this notebook.

", "rawMarkdown": "Creative approach! Just a word of caution that testing ways to hack the API in a real submission may look sufficiently like a cheating attempt to us that we end up making a ban at the end of the competition. That's true regardless of whether or not the method actually ends up working. \n\nWith that said, @yihdarshieh you don't need to worry about getting banned for this notebook.", "votes": null }, { "id": "1049902", "postDate": "10/14/2020 21:52:30", "content": "

@sohier Thanks for being kind to me. However, I thought that if we don't select these hacking notebooks for final 2 submissions, we should be fine. Is the rule so strict that it will ban even we don't use it for private LB score?

", "rawMarkdown": "sohier Thanks for being kind to me. However, I thought that if we don't select these hacking notebooks for final 2 submissions, we should be fine. Is the rule so strict that it will ban even we don't use it for private LB score?", "votes": null }, { "id": "1049922", "postDate": "10/14/2020 22:13:34", "content": "

We won't ignore hacking attempts just because they weren't selected.

", "rawMarkdown": "We won't ignore hacking attempts just because they weren't selected.", "votes": null } ], "comments": [ { "id": 1049894, "author_name": "dwit392", "author_url": "", "post_date": "10/14/2020 21:38:26", "content": "

I have a question for you. When your notebook failed after 9 hours, did you receive a submission scoring error or a different error?

", "votes": null, "replies": [ { "id": 1049895, "author_name": "yihdarshieh", "author_url": "", "post_date": "10/14/2020 21:39:38", "content": "

timeout error

", "votes": null, "replies": [] } ] }, { "id": 1049897, "author_name": "sohier", "author_url": "", "post_date": "10/14/2020 21:48:48", "content": "

Creative approach! Just a word of caution that testing ways to hack the API in a real submission may look sufficiently like a cheating attempt to us that we end up making a ban at the end of the competition. That's true regardless of whether or not the method actually ends up working.

\n

With that said, @yihdarshieh you don't need to worry about getting banned for this notebook.

", "votes": null, "replies": [ { "id": 1049902, "author_name": "yihdarshieh", "author_url": "", "post_date": "10/14/2020 21:52:30", "content": "

@sohier Thanks for being kind to me. However, I thought that if we don't select these hacking notebooks for final 2 submissions, we should be fine. Is the rule so strict that it will ban even we don't use it for private LB score?

", "votes": null, "replies": [] }, { "id": 1049922, "author_name": "sohier", "author_url": "", "post_date": "10/14/2020 22:13:34", "content": "

We won't ignore hacking attempts just because they weren't selected.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1049864": "My notebook [https://www.kaggle.com/yihdarshieh/able-to-call-make-env-multiple-times/](https://www.kaggle.com/yihdarshieh/able-to-call-make-env-multiple-times/) shows how to call `make_env()` multiple times.\n\n**IMPORTANT** This only works in interaction mode or commit mode. Submission will fail.\n\nThis notebook shows you how to call make_env() multiple times in the same notebook session. It works also when we commit the notebook. However, the submission will fail (it will run for 9 hours and timeout).\n\nBut while you build the input pipeline or models, this might make your work easier (without restarting the notebook).\n\nI hope this hack won't never work for submission! If someone smart enough (not me 😓) to do it, i hope he/she will contact the host.", "1049894": "I have a question for you. When your notebook failed after 9 hours, did you receive a submission scoring error or a different error?", "1049895": "timeout error", "1049897": "Creative approach! Just a word of caution that testing ways to hack the API in a real submission may look sufficiently like a cheating attempt to us that we end up making a ban at the end of the competition. That's true regardless of whether or not the method actually ends up working. \n\nWith that said, @yihdarshieh you don't need to worry about getting banned for this notebook.", "1049902": "sohier Thanks for being kind to me. However, I thought that if we don't select these hacking notebooks for final 2 submissions, we should be fine. Is the rule so strict that it will ban even we don't use it for private LB score?", "1049922": "We won't ignore hacking attempts just because they weren't selected." }, "source": "meta" }