There are some reports that submissions became to be getting low scores while it scored well before API update.
\nI also met the same problem today.
\nIs there anyone who get healthy score recently?
I think sharing the recent status of the submission health publicly will help people to know whether they have problem in their own code or in the API.
\nI want you to share the information about
\nThank you!
", "messages": [ { "id": "2285898", "postDate": "06/03/2023 04:57:28", "content": "There are some reports that submissions became to be getting low scores while it scored well before API update.
\nI also met the same problem today.
\nIs there anyone who get healthy score recently?
I think sharing the recent status of the submission health publicly will help people to know whether they have problem in their own code or in the API.
\nI want you to share the information about
\nThank you!
", "rawMarkdown": "There are some reports that submissions became to be getting low scores while it scored well before API update.\n\n- https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414109#2282932\n\n- https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414109#2283373\n\nI also met the same problem today.\nIs there anyone who get healthy score recently?\n\nI think sharing the recent status of the submission health publicly will help people to know whether they have problem in their own code or in the API.\n\nI want you to share the information about\n\n1. the date you submitted\n2. whether the submission succeeded, failed, or unhealthy(low scored)\n(3. If you have a problem and successfully solved it, I want you to share the cause and the solution.)\n\nThank you!", "votes": null }, { "id": "2285900", "postDate": "06/03/2023 05:00:16", "content": "1: 6/3
\n2: unhealthy (previously LB: 0.695 notebook got LB: 563)
f**k game， waste all my time！
", "rawMarkdown": "f**k game， waste all my time！", "votes": null }, { "id": "2286069", "postDate": "06/03/2023 07:46:43", "content": "sadge.
\nthanks for sharing.
My recent submissions have passed with score ~0.69x (most recent Jun 3 11:40 AM UTC)
\nI think these steps are important to avoid getting low scores (i.e. sort test by index and sort questions also)
\n (test, sample_submission) iter_test:\n test = test(by = )\n sample_submission = sample_submission(lambda x: x())(int)\n sample_submission = sample_submission(by = )\n sample_submission(,axis=,inplace=True)\nMay I ask if you are using the old API or the new one？
", "rawMarkdown": "May I ask if you are using the old API or the new one？", "votes": null }, { "id": "2286483", "postDate": "06/03/2023 13:51:52", "content": "I tried with both old and new (i.e. import jo_wilder and import jo_wilder_310 ) - same score and same inference time
Hi @rashmibanthia,
\nIf we need to sort test DataFrame by index to somewhat recover the ordering of events, I think there still exist issues related to the intrinsic data error within theindex column. As pointed out here, index has problem like reversed index. Hence, the manual sorting may mess up the real ordering of events. But, this can be a temporary workaround. Thanks for your sharing!
yes, I noticed that in your training kernel as well, more specific this - drop_multi_game_naive()
\nI actually disabled that while training also. it's likely that I missed something considering my score and I joined this competition quite late
1: 6/4
\n2: succeeded
\n3:
\nproblem:
\nhttps://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414751#2285900
cause: The order of the data given by the API is not ordered.
\n\n@rashmibanthia
\nThanks a lot for your sharing!
Need Help
\nI got an issue when trying to submit my code to Kaggle with an error message stating \"Notebook Threw Exception.\" When I tested the code in Jupyter on Kaggle, everything worked fine.
\nHere is the link to my Kaggle notebook:
\nhttps://www.kaggle.com/code/lintanggilangpratama/why-can-t-submission/notebook.
", "rawMarkdown": "Need Help\n\nI got an issue when trying to submit my code to Kaggle with an error message stating \"Notebook Threw Exception.\" When I tested the code in Jupyter on Kaggle, everything worked fine. \n\nHere is the link to my Kaggle notebook: \n\nhttps://www.kaggle.com/code/lintanggilangpratama/why-can-t-submission/notebook.", "votes": null }, { "id": "2288029", "postDate": "06/05/2023 06:11:50", "content": "I submitted 6/1, 6/2, 6/4, no error thus far.
\nAfter sorting by index, old submission result in < 0.001 reduction in LB score (Different order when sorting the submissions by public score).
However as @abaojiang mentioned in this post: https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/395250, the index in its natural order doesn't always increase monotonically, which might explain the LB difference I'm seeing.
", "rawMarkdown": "I submitted 6/1, 6/2, 6/4, no error thus far. \nAfter sorting by index, old submission result in < 0.001 reduction in LB score (Different order when sorting the submissions by public score). \n\nHowever as @abaojiang mentioned in this post: https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/395250, the index in its natural order doesn't always increase monotonically, which might explain the LB difference I'm seeing.", "votes": null } ], "comments": [ { "id": 2285900, "author_name": "nynyny67", "author_url": "", "post_date": "06/03/2023 05:00:16", "content": "1: 6/3
\n2: unhealthy (previously LB: 0.695 notebook got LB: 563)
f**k game， waste all my time！
", "votes": null, "replies": [ { "id": 2286069, "author_name": "nynyny67", "author_url": "", "post_date": "06/03/2023 07:46:43", "content": "sadge.
\nthanks for sharing.
My recent submissions have passed with score ~0.69x (most recent Jun 3 11:40 AM UTC)
\nI think these steps are important to avoid getting low scores (i.e. sort test by index and sort questions also)
\n (test, sample_submission) iter_test:\n test = test(by = )\n sample_submission = sample_submission(lambda x: x())(int)\n sample_submission = sample_submission(by = )\n sample_submission(,axis=,inplace=True)\nMay I ask if you are using the old API or the new one？
", "votes": null, "replies": [ { "id": 2286483, "author_name": "rashmibanthia", "author_url": "", "post_date": "06/03/2023 13:51:52", "content": "I tried with both old and new (i.e. import jo_wilder and import jo_wilder_310 ) - same score and same inference time
Hi @rashmibanthia,
\nIf we need to sort test DataFrame by index to somewhat recover the ordering of events, I think there still exist issues related to the intrinsic data error within theindex column. As pointed out here, index has problem like reversed index. Hence, the manual sorting may mess up the real ordering of events. But, this can be a temporary workaround. Thanks for your sharing!
yes, I noticed that in your training kernel as well, more specific this - drop_multi_game_naive()
\nI actually disabled that while training also. it's likely that I missed something considering my score and I joined this competition quite late
1: 6/4
\n2: succeeded
\n3:
\nproblem:
\nhttps://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414751#2285900
cause: The order of the data given by the API is not ordered.
\n\n@rashmibanthia
\nThanks a lot for your sharing!
Need Help
\nI got an issue when trying to submit my code to Kaggle with an error message stating \"Notebook Threw Exception.\" When I tested the code in Jupyter on Kaggle, everything worked fine.
\nHere is the link to my Kaggle notebook:
\nhttps://www.kaggle.com/code/lintanggilangpratama/why-can-t-submission/notebook.
", "votes": null, "replies": [] }, { "id": 2288029, "author_name": "aesoptacit", "author_url": "", "post_date": "06/05/2023 06:11:50", "content": "I submitted 6/1, 6/2, 6/4, no error thus far.
\nAfter sorting by index, old submission result in < 0.001 reduction in LB score (Different order when sorting the submissions by public score).
However as @abaojiang mentioned in this post: https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/395250, the index in its natural order doesn't always increase monotonically, which might explain the LB difference I'm seeing.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2285898": "There are some reports that submissions became to be getting low scores while it scored well before API update.\n\n- https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414109#2282932\n\n- https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414109#2283373\n\nI also met the same problem today.\nIs there anyone who get healthy score recently?\n\nI think sharing the recent status of the submission health publicly will help people to know whether they have problem in their own code or in the API.\n\nI want you to share the information about\n\n1. the date you submitted\n2. whether the submission succeeded, failed, or unhealthy(low scored)\n(3. If you have a problem and successfully solved it, I want you to share the cause and the solution.)\n\nThank you!", "2285900": "1: 6/3\n2: unhealthy (previously LB: 0.695 notebook got LB: 563)", "2286061": "1. 5/31 - 6/3\n2. all notebook submission score error! (11 submission with 3 of them submitted before and get score successfully)", "2286063": "f**k game， waste all my time！", "2286069": "sadge.\nthanks for sharing.", "2286322": "My recent submissions have passed with score ~0.69x (most recent Jun 3 11:40 AM UTC) \n\nI think these steps are important to avoid getting low scores (i.e. sort test by index and sort questions also)\n\n```\nfor (test, sample_submission) in iter_test:\n test = test.sort_values(by = 'index')\n sample_submission['q'] = sample_submission.session_id.apply(lambda x: x.split(\"_q\")[1]).astype(int)\n sample_submission = sample_submission.sort_values(by = 'q')\n sample_submission.drop('q',axis=1,inplace=True)\n```", "2286460": "May I ask if you are using the old API or the new one？", "2286483": "I tried with both old and new (i.e. `import jo_wilder` and `import jo_wilder_310` ) - same score and same inference time", "2286528": "Hi @rashmibanthia,\n\nIf we need to sort test DataFrame by `index` to somewhat recover the ordering of events, I think there still exist issues related to the intrinsic data error within the`index` column. As pointed out [here](https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/395250), `index` has problem like **reversed index**. Hence, the manual sorting may mess up the real ordering of events. But, this can be a temporary workaround. Thanks for your sharing!", "2286539": "yes, I noticed that in your training kernel as well, more specific [this](https://www.kaggle.com/code/abaojiang/lb-0-694-tconv-with-4-features-training-part?scriptVersionId=124329755&cellId=7) - `drop_multi_game_naive()`\nI actually disabled that while training also. it's likely that I missed something considering my score and I joined this competition quite late", "2286829": "1: 6/4\n2: succeeded\n3: \nproblem:\nhttps://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414751#2285900\n\ncause: The order of the data given by the API is not ordered.\n\nsolution: https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/414751#2286322)\n\n@rashmibanthia \nThanks a lot for your sharing!", "2287043": "Need Help\n\nI got an issue when trying to submit my code to Kaggle with an error message stating \"Notebook Threw Exception.\" When I tested the code in Jupyter on Kaggle, everything worked fine. \n\nHere is the link to my Kaggle notebook: \n\nhttps://www.kaggle.com/code/lintanggilangpratama/why-can-t-submission/notebook.", "2288029": "1. 6/5\n2. not sure...\nUsing RB's code resulted in a score close to the previous one, but not exactly the same.", "2288271": "I submitted 6/1, 6/2, 6/4, no error thus far. \nAfter sorting by index, old submission result in < 0.001 reduction in LB score (Different order when sorting the submissions by public score). \n\nHowever as @abaojiang mentioned in this post: https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/395250, the index in its natural order doesn't always increase monotonically, which might explain the LB difference I'm seeing." }, "source": "meta" }