We currently support fp16 (half) codebook vectors types for PQ dataset and Iterative CAGRA-Q build.
The CAGRA-Q search kernels already support float codebook vectors types as well. We can explore whether it is worth implementing the float support end-to-end and if that has any benefits.
#1810 (comment)
We currently support fp16 (half) codebook vectors types for PQ dataset and Iterative CAGRA-Q build.
The CAGRA-Q search kernels already support float codebook vectors types as well. We can explore whether it is worth implementing the float support end-to-end and if that has any benefits.
#1810 (comment)