deepseek-v4.1POST
vLLM + EXL3 量化本地跑 DeepSeek V4.1 Flash 1017 仅 1 tok/s
用户用 ViC305 量化配 vllm-exl3 部署 DeepSeek V4.1 Flash 1017，几小时后模型加载稳定，但速度只有 1 tok/s，正在排查根因。
Hassan Emam
@hassanemam9 天前收录原推发布于 2026.10.23
Hassan Emam @hassanemam原推发布于 2026.10.23
I am experimenting serving deepseek V4.1 Flash 1017 using @ViC305 quantization and vllm-exl3. After few hours I managed to load the model and make it stable. The bad news it is running on 1 tok/s. I am still working on finding the root cause. Stay tunned
x.com/hassanemam/status/2113325247533363247采集于 2026.10.23 12:08