Repository navigation
Add MicroKernels to support f32, f16, u8, i8 and MinMax quantizations. #990
Description
Add MicroKernels for required architecture-type/quantization pairs (Blanket Issue)
Description
Implementation of the hardware-specific microkernels for multi-vector reductions.
Tasks
- Implement microkernel for different types and quantizations for multi-vector distances (Chamfer/MaxSim).
- Ensure the microkernel logic is decoupled enough to be tested via the standalone API.
- Blanket Task: Add subsequent sub-issues for ARM NEON, AVX-512, etc., as needed.
Metadata
Metadata
Assignees
Labels
Type
Projects
- Status
Activity
[-]Add MicroKernels for required architecture-type/quantization pairs[/-][+]Add MicroKernels for required architecture-type/quantization pairs (Blanket Issue)[/+][-]Add MicroKernels for required architecture-type/quantization pairs (Blanket Issue)[/-][+]Add MicroKernels for required architecture-type/quantization pairs[/+][-]Add MicroKernels for required architecture-type/quantization pairs[/-][+]Add MicroKernels to support f32, f16, u8, i8 and MinMax quantizations.[/+]