TurboQuant can reduce vector index size by 10x at 100M Row Scale(github.com)

8 pointsby mxfeinberg9 dni temu2 comments

https://github.com/pgvector/pgvector/pull/989

3 comments

0-_-08 dni temu

32 bits vs 4 bits it looks like

Yup, and unlike the original turboquant paper, my implementation is pinned to using a 4 bit code book so I could use SIMD kernels for performance.

Slovian8 dni temu