https://pypi.org/project/vllm-xtu-moe/ Recent updates to the Python Package Index for vllm-xtu-moe en https://pypi.org/project/vllm-xtu-moe/0.2.7/ XTU (X Transformers Unity) — a mainline-vLLM hybrid MoE plugin (CPU experts + GPU attention/long prefill). Bundles the xiaotu CPU MoE engine (AVX512-VNNI/BF16: MXFP4/FP8/INT4/BF16) so no-AMX x86 hosts (AMD EPYC) get usable CPU MoE backends through FusedMoEFactory. dahema@me.com Fri, 09 Oct 2026 09:46:22 GMT