llama.cpp b11856：新增投机解码拒绝采样测试

AI 摘要

llama.cpp 发布 b11856 版本，主要变更是新增投机解码拒绝采样（speculative rejection sampling）的测试（#30213），用于验证草稿模型产生的 token 与目标模型分布一致，确保拒绝采样后仍保持与直接采样目标模型相同的概率分布。该版本同时提供 macOS、Linux、Windows、Android 等多平台预编译包，覆盖 CUDA、Vulkan、ROCm、SYCL、OpenVINO 等后端。

正文节选

<details open> tests : add speculative rejection draft sampling test (#30213) * tests : add speculative rejection draft sampling test This commit adds a test for speculative rejection sampling test that excercises the speculative rejection sampling mechanism in common/speculative.cpp. The motivation for this is have a test that verifies that the speculative draft models emitts tokens that match the target models distribution. So when rejection sampling occurs it should still produce the sam

发布时间：2026-10-24 23:06

抓取时间：2026-10-24 23:37

来源机构：ggml-org