LMSYS：Blog（Chatbot Arena 团队）·· 13 小时前AI 评分46
SGLang 推出 EPD 分离架构：为视觉语言模型实现弹性编码器扩展
Blog EPD Disaggregation: Elastic Encoder Scaling for Vision-Language Models in SGLang We introduce Encoder-Prefill-Decode (EPD) Disaggregation in SGLang, a novel architecture that separates vision encoding from language processing in Vision-Language Models (VLMs). This can enable: - I... rednote hilab, Alibaba Cloud Computing, AntGroup SCT January 12, 2026
AI 导读
SGLang 推出 Encoder-Prefill-Decode（EPD）分离架构，将视觉语言模型（VLM）的视觉编码与语言处理拆分为三层，使编码器服务器可独立横向扩展。
来源：LMSYS：Blog（Chatbot Arena 团队） · lmsys.org