v0.18.1 | FList

名称

大小

更新时间

vllm-0.18.1-cp38-abi3-manylinux_2_31_aarch64.whl

2026-03-31 00:54

vllm-0.18.1-cp38-abi3-manylinux_2_31_x86_64.whl

2026-03-31 00:54

vllm-0.18.1+cu130-cp38-abi3-manylinux_2_35_aarch64.whl

2026-03-31 00:54

vllm-0.18.1+cu130-cp38-abi3-manylinux_2_35_x86_64.whl

2026-03-31 00:54

vllm-0.18.1.tar.gz

2026-03-31 00:53

vllm-0.18.1+cpu-cp38-abi3-manylinux_2_35_x86_64.whl

2026-03-31 00:53

vllm-0.18.1+cpu-cp38-abi3-manylinux_2_35_aarch64.whl

2026-03-31 00:53

This is a patch release on top of v0.18.0 to address a few issues:

Change default SM100 MLA prefill backend back to TRT-LLM (#38562)
Fix mock.patch resolution failure for standalone_compile.FakeTensorMode on Python <= 3.10 (#37158)
Disable monolithic TRTLLM MoE for Renormalize routing #37605
Pre-download missing FlashInfer headers in Docker build #38391
Fix DeepGemm E8M0 accuracy degradation for Qwen3.5 FP8 on Blackwell (#38083)