Open Models Almanac
Methodsinference & servinginference quantization

single mention · inference & serving

FP8 or NVFP4 KV-cache quantization

A KV-cache configuration that quantizes both key and value caches in either FP8 or NVFP4.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See inference quantization for the methods filed beside it.

Also called FP8 and NVFP4 quantization cases.

Filed under inference & serving › inference quantization