I have forked a notebook with no changes but on submission the old one is perfect and yielded a submission score but the new forked one threw notebook timeout error. Can anyone please explain this???
", "messages": [ { "id": 1061186, "postDate": "2020-10-26T19:40:00.197Z", "content": "I have forked a notebook with no changes but on submission the old one is perfect and yielded a submission score but the new forked one threw notebook timeout error. Can anyone please explain this???
", "rawMarkdown": "I have forked a notebook with no changes but on submission the old one is perfect and yielded a submission score but the new forked one threw notebook timeout error. Can anyone please explain this???", "votes": 5 }, { "id": 1061189, "postDate": "2020-10-26T19:44:29.600Z", "content": "The broader picture is discussed here: https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/discussion/193252
\nEssentially, the hypothesis is that different CPU's are used resulting in different I/O times, causing some submissions to run and others cannot. In addition, there was a bit of an outage this morning. However, according to a Kaggle team member,
\n\n", "rawMarkdown": "The broader picture is discussed here: https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/discussion/193252\n\nEssentially, the hypothesis is that different CPU's are used resulting in different I/O times, causing some submissions to run and others cannot. In addition, there was a bit of an outage this morning. However, according to a Kaggle team member,\n\n> This did not affect commit runtime/timeouts at all, it just meant your commit might have taken longer to be assigned a machine to execute it.\n", "votes": 1 }, { "id": 1061235, "postDate": "2020-10-26T20:51:53.660Z", "content": "This did not affect commit runtime/timeouts at all, it just meant your commit might have taken longer to be assigned a machine to execute it.
\n
same here, it's taking longer than expected.
", "rawMarkdown": "same here, it's taking longer than expected." }, { "id": 1061193, "postDate": "2020-10-26T19:45:45.327Z", "content": "Yes same happened to us. Our 2 final submission were from the same notebook, same code only difference is some code was commented to use 1 less model. The submission with 1 less model failed and the other one with full number of models passed. This shows inconsistency and unreliability of Kaggle platform. Also one time same submission ran in 7 hours in other time 9 hours.
", "rawMarkdown": "Yes same happened to us. Our 2 final submission were from the same notebook, same code only difference is some code was commented to use 1 less model. The submission with 1 less model failed and the other one with full number of models passed. This shows inconsistency and unreliability of Kaggle platform. Also one time same submission ran in 7 hours in other time 9 hours." } ], "comments": [ { "id": 1061189, "author_name": "Stanley Zheng", "author_url": "", "post_date": "2020-10-26T19:44:29.600000", "content": "The broader picture is discussed here: https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/discussion/193252
\nEssentially, the hypothesis is that different CPU's are used resulting in different I/O times, causing some submissions to run and others cannot. In addition, there was a bit of an outage this morning. However, according to a Kaggle team member,
\n\n", "votes": 1, "replies": [] }, { "id": 1061235, "author_name": "sin", "author_url": "", "post_date": "2020-10-26T20:51:53.660000", "content": "This did not affect commit runtime/timeouts at all, it just meant your commit might have taken longer to be assigned a machine to execute it.
\n
same here, it's taking longer than expected.
", "votes": 0, "replies": [] }, { "id": 1061193, "author_name": "Kerem Turgutlu", "author_url": "", "post_date": "2020-10-26T19:45:45.327000", "content": "Yes same happened to us. Our 2 final submission were from the same notebook, same code only difference is some code was commented to use 1 less model. The submission with 1 less model failed and the other one with full number of models passed. This shows inconsistency and unreliability of Kaggle platform. Also one time same submission ran in 7 hours in other time 9 hours.
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "1061186": "I have forked a notebook with no changes but on submission the old one is perfect and yielded a submission score but the new forked one threw notebook timeout error. Can anyone please explain this???", "1061189": "The broader picture is discussed here: https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/discussion/193252\n\nEssentially, the hypothesis is that different CPU's are used resulting in different I/O times, causing some submissions to run and others cannot. In addition, there was a bit of an outage this morning. However, according to a Kaggle team member,\n\n> This did not affect commit runtime/timeouts at all, it just meant your commit might have taken longer to be assigned a machine to execute it.\n", "1061235": "same here, it's taking longer than expected.", "1061193": "Yes same happened to us. Our 2 final submission were from the same notebook, same code only difference is some code was commented to use 1 less model. The submission with 1 less model failed and the other one with full number of models passed. This shows inconsistency and unreliability of Kaggle platform. Also one time same submission ran in 7 hours in other time 9 hours." } }