feat: add per-tensor FP8 e4m3 quant/dequant ops for KV cache quantiza… - #1521

Open

BoBoDai wants to merge 2 commits into

InfiniTensor:mainfrom

BoBoDai:suppose-fp8

Commits on Aug 23, 2026

feat: add per-tensor FP8 e4m3 quant/dequant ops for KV cache quantization
BoBoDai
committed
e4c34c7
fix: guard per-tensor FP8 quant against zero scale (all-zero input)
BoBoDai
committed
036fb70