top | item 16711933 Making a faster K-NN algorithm 5 points| samgrassi | 8 years ago |medium.com | reply 1 comment order hn newest [+] [-] TheAnig|8 years ago|reply Pretty neat read. Cool to see that something as trivial as the distance measure, can save up on FLOPs so much that it beats scikit's default.
[+] [-] TheAnig|8 years ago|reply Pretty neat read. Cool to see that something as trivial as the distance measure, can save up on FLOPs so much that it beats scikit's default.
[+] [-] TheAnig|8 years ago|reply