{ "id": 472916, "title": "the '900' magic ", "url": "/competitions/blood-vessel-segmentation/discussion/472916", "author_name": "", "post_date": "2024-02-02T17:11:13.511960600Z", "votes": null, "comment_count": 2, "views": 0, "content": "

I thought I read all the discussion threads but I probably missed this.
\nThe post(s) I saw it was not citing any apparent reasoning.
\nThen I tried a simple search on competition discussion forum for the key: '900', and as expected the super-vintage Kaggle search did not reveal anything.
\nSo what is this magic number for kidney-2 about: training/validation on slices 900+?

", "messages": [ { "id": "2632866", "postDate": "02/02/2024 17:11:13", "content": "

I thought I read all the discussion threads but I probably missed this.
\nThe post(s) I saw it was not citing any apparent reasoning.
\nThen I tried a simple search on competition discussion forum for the key: '900', and as expected the super-vintage Kaggle search did not reveal anything.
\nSo what is this magic number for kidney-2 about: training/validation on slices 900+?

", "rawMarkdown": "I thought I read all the discussion threads but I probably missed this. \nThe post(s) I saw it was not citing any apparent reasoning.\nThen I tried a simple search on competition discussion forum for the key: '900', and as expected the super-vintage Kaggle search did not reveal anything. \nSo what is this magic number for kidney-2 about: training/validation on slices 900+?", "votes": null }, { "id": "2633112", "postDate": "02/02/2024 20:05:52", "content": "

First mention of this shift in annotation by @hengck23 :
\nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2543380

\n

where the 900 number came from:
\nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2589018

\n
\n

i use train= kidney1 dense + kidney3 dense
\n validation= kidney2 sparse (from slice >= 900 to avoid annotation shift error)

\n
", "rawMarkdown": "First mention of this shift in annotation by @hengck23 : \nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2543380\n\nwhere the 900 number came from:\nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2589018\n\n> i use train= kidney1 dense + kidney3 dense\nvalidation= kidney2 sparse (from slice >= 900 to avoid annotation shift error)", "votes": null }, { "id": "2633189", "postDate": "02/02/2024 21:05:25", "content": "

that is helpful thank you.

", "rawMarkdown": "that is helpful thank you.", "votes": null } ], "comments": [ { "id": 2633112, "author_name": "janmpia", "author_url": "", "post_date": "02/02/2024 20:05:52", "content": "

First mention of this shift in annotation by @hengck23 :
\nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2543380

\n

where the 900 number came from:
\nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2589018

\n
\n

i use train= kidney1 dense + kidney3 dense
\n validation= kidney2 sparse (from slice >= 900 to avoid annotation shift error)

\n
", "votes": null, "replies": [ { "id": 2633189, "author_name": "abdulkadirguner", "author_url": "", "post_date": "02/02/2024 21:05:25", "content": "

that is helpful thank you.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2632866": "I thought I read all the discussion threads but I probably missed this. \nThe post(s) I saw it was not citing any apparent reasoning.\nThen I tried a simple search on competition discussion forum for the key: '900', and as expected the super-vintage Kaggle search did not reveal anything. \nSo what is this magic number for kidney-2 about: training/validation on slices 900+?", "2633112": "First mention of this shift in annotation by @hengck23 : \nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2543380\n\nwhere the 900 number came from:\nhttps://www.kaggle.com/competitions/blood-vessel-segmentation/discussion/456118#2589018\n\n> i use train= kidney1 dense + kidney3 dense\nvalidation= kidney2 sparse (from slice >= 900 to avoid annotation shift error)", "2633189": "that is helpful thank you." }, "source": "meta" }