deepseek-v4.1IMAGE
vLLM v0.32.0: ~60% DeepSeek-V4.1 throughput gain on H20
vLLM v0.32.0 ships DaoCloud's H20 inference optimizations, reporting roughly 60% higher DeepSeek-V4.1 throughput.
paco xu
@xu_pacoArchived 3 days agoOriginally posted 2026.10.26
paco xu @xu_pacoOriginally posted 2026.10.26
vLLM v0.32.0 发布：DaoCloud 优化 H20 推理，DeepSeek-V4.1 吞吐提升约 60% https://mp.weixin.qq.com/s/Jhd-fiW-uKT8SY0tvTMTCA
This post quotes another post: x.com/vllm_project/status/2112760256120197333
x.com/xu_paco/status/2114555986463908265Archived 2026.10.26 12:02