![\"notebook](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7410561%2Fa64348d39a3368158b901752b93cdd81%2FScreenshot%20from%202023-04-14%2016-24-01.png?generation=1681457084513937&alt=media\")
Due to some issues with original example submission kernel https://www.kaggle.com/code/eduardtrulls/imc-2023-submission-example, we have updated it:
\nWe apologize for the inconveniences with the first version.
\nHappy kaggling!
", "messages": [ { "id": 2221001, "postDate": "2023-04-13T21:28:10.187Z", "content": "Due to some issues with original example submission kernel https://www.kaggle.com/code/eduardtrulls/imc-2023-submission-example, we have updated it:
\nWe apologize for the inconveniences with the first version.
\nHappy kaggling!
", "rawMarkdown": "Due to some issues with original example submission kernel https://www.kaggle.com/code/eduardtrulls/imc-2023-submission-example, we have updated it:\n\n- fixed a bug in picking the best reconstruction\n- decreased number of features to 2048, so it should finish in 2 and not 8 hours. \n\nWe apologize for the inconveniences with the first version. \n\nHappy kaggling!", "votes": 8 }, { "id": 2221343, "postDate": "2023-04-14T07:27:54.327Z", "content": "I think you guys have no complete(full) version of sample_submission.csv in the submission environment. Here is what I got from the given example submission notebook. I guess that explains the weird score distribution on the current leader board?
\n
This is standard for the Code competitions. When you click \"Submit\", the dummy sample_submission.csv is replaced with real one, not shown to user. In other words - no, this doesn't explain the scores. It is just that datasets are hard and hyper parameters are not tuned for quality in sample submission.
You are right..Thank you for the answer tho
", "rawMarkdown": "You are right..Thank you for the answer tho" }, { "id": 2221460, "postDate": "2023-04-14T09:37:35.513Z", "content": "I came up with this doubt because the submission took no longer than a minute which is different from what’s mentioned here. Maybe there’s something I’ve done wrong
", "rawMarkdown": "I came up with this doubt because the submission took no longer than a minute which is different from what’s mentioned here. Maybe there’s something I’ve done wrong" }, { "id": 2221547, "postDate": "2023-04-14T10:56:07.823Z", "content": "Can it be that not submission, but saving took no longer than a minute?
", "rawMarkdown": "Can it be that not submission, but saving took no longer than a minute?" }, { "id": 2221656, "postDate": "2023-04-14T12:44:01.630Z", "content": "You might be running the notebook without submitting, in which case the test data is not populated and there's nothing to process. You have to click the \"submit\" button (after disabling internet access) to actually queue your submission.
", "rawMarkdown": "You might be running the notebook without submitting, in which case the test data is not populated and there's nothing to process. You have to click the \"submit\" button (after disabling internet access) to actually queue your submission.", "votes": 1 }, { "id": 2221745, "postDate": "2023-04-14T14:35:14.720Z", "content": "Sorry for bothering you guys. Maybe I'm doing something terribly wrong. But I clicked \"submit\" and got this in a minute or so. I think that several people on the leaderboard might have run the same notebook and got the same result.
\n![\"submit\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7410561%2F50e8ec104bae54a0a1b9380c79ab1254%2FScreenshot%20from%202023-04-14%2023-14-18.png?generation=1681482594375160&alt=media\")
No worries. I queued the last version of the notebook and it's been running for longer than that:
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F9806603%2F63d111458e1670f40b6637eea6d3b3fc%2FScreenshot%202023-04-14%20at%2017.11.11.png?generation=1681485096175990&alt=media\")
You seem to be doing it correctly though, as far as I can tell. Are you using the last version?
", "rawMarkdown": "No worries. I queued the last version of the notebook and it's been running for longer than that:\n\n\n\nYou seem to be doing it correctly though, as far as I can tell. Are you using the last version?", "votes": 1 }, { "id": 2221803, "postDate": "2023-04-14T15:30:45.237Z", "content": "Yes. The website did not show me the real-time running state of the notebook so I refreshed the website, and saw that result. Did you refresh the website and is the notebook still running?
", "rawMarkdown": "Yes. The website did not show me the real-time running state of the notebook so I refreshed the website, and saw that result. Did you refresh the website and is the notebook still running? " }, { "id": 2221844, "postDate": "2023-04-14T16:19:20.643Z", "content": "Yes, it's been running for 1h.
", "rawMarkdown": "Yes, it's been running for 1h." }, { "id": 2224072, "postDate": "2023-04-17T01:05:07.723Z", "rawMarkdown": "", "votes": 1, "isDeleted": true } ] } ] } ], "comments": [ { "id": 2221343, "author_name": "Hyeokjoon Kwon (권혁준)", "author_url": "", "post_date": "2023-04-14T07:27:54.327000", "content": "I think you guys have no complete(full) version of sample_submission.csv in the submission environment. Here is what I got from the given example submission notebook. I guess that explains the weird score distribution on the current leader board?
\n
This is standard for the Code competitions. When you click \"Submit\", the dummy sample_submission.csv is replaced with real one, not shown to user. In other words - no, this doesn't explain the scores. It is just that datasets are hard and hyper parameters are not tuned for quality in sample submission.
You are right..Thank you for the answer tho
", "votes": 0, "replies": [] }, { "id": 2221460, "author_name": "Hyeokjoon Kwon (권혁준)", "author_url": "", "post_date": "2023-04-14T09:37:35.513000", "content": "I came up with this doubt because the submission took no longer than a minute which is different from what’s mentioned here. Maybe there’s something I’ve done wrong
", "votes": 0, "replies": [] }, { "id": 2221547, "author_name": "old-ufo", "author_url": "", "post_date": "2023-04-14T10:56:07.823000", "content": "Can it be that not submission, but saving took no longer than a minute?
", "votes": 0, "replies": [] }, { "id": 2221656, "author_name": "Eduard Trulls", "author_url": "", "post_date": "2023-04-14T12:44:01.630000", "content": "You might be running the notebook without submitting, in which case the test data is not populated and there's nothing to process. You have to click the \"submit\" button (after disabling internet access) to actually queue your submission.
", "votes": 1, "replies": [] }, { "id": 2221745, "author_name": "Hyeokjoon Kwon (권혁준)", "author_url": "", "post_date": "2023-04-14T14:35:14.720000", "content": "Sorry for bothering you guys. Maybe I'm doing something terribly wrong. But I clicked \"submit\" and got this in a minute or so. I think that several people on the leaderboard might have run the same notebook and got the same result.
\n
No worries. I queued the last version of the notebook and it's been running for longer than that:
\n
You seem to be doing it correctly though, as far as I can tell. Are you using the last version?
", "votes": 1, "replies": [] }, { "id": 2221803, "author_name": "Hyeokjoon Kwon (권혁준)", "author_url": "", "post_date": "2023-04-14T15:30:45.237000", "content": "Yes. The website did not show me the real-time running state of the notebook so I refreshed the website, and saw that result. Did you refresh the website and is the notebook still running?
", "votes": 0, "replies": [] }, { "id": 2221844, "author_name": "Eduard Trulls", "author_url": "", "post_date": "2023-04-14T16:19:20.643000", "content": "Yes, it's been running for 1h.
", "votes": 0, "replies": [] }, { "id": 2224072, "author_name": "", "author_url": "", "post_date": "2023-04-17T01:05:07.723000", "content": "", "votes": 1, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2221001": "Due to some issues with original example submission kernel https://www.kaggle.com/code/eduardtrulls/imc-2023-submission-example, we have updated it:\n\n- fixed a bug in picking the best reconstruction\n- decreased number of features to 2048, so it should finish in 2 and not 8 hours. \n\nWe apologize for the inconveniences with the first version. \n\nHappy kaggling!", "2221343": "I think you guys have no complete(full) version of sample_submission.csv in the submission environment. Here is what I got from the given example submission notebook. I guess that explains the weird score distribution on the current leader board? \n\n" } }