Ant Group and vLLM open Humming for quantized MoE and dense inference
Unified quantization IR, JIT kernels, Turing through Blackwell. Serving-stack follow-on for quantized GEMM paths.
Read the piece#kernels
One published Frontier piece on this subject — open it below.
1 article · Full archive · Topics
Read this piece
This tag currently points at a single story.
One brief a week.
New releases, then the constraint.