llama.cpp b11922：修复多序列拆分回放测试

AI 摘要

llama.cpp 发布 b11922 版本，主要修复了多序列（multi-seq）拆分回放测试中针对已训练模型的若干问题：调整回滚快照与参考上下文的比较方式，仅比较每个序列首个回放 token，并要求 seq-1 隔离达到位精确；同时移除基于 llama_model_n_swa() 的部分状态平坦性检查，保留 partial <= full 不变量。该版本同步提供 macOS、Linux、Windows、Android 等多平台预编译包。

正文节选

<details open> tests : fix multi-seq split replay and partial-state cases for trained models (#31216) * tests : fix multi-seq split replay against trained models The rollback snapshot is captured by the graph that decoded the rolled-back tail, while the reference context decodes the same prefix with different batch shapes. The two states differ in the last bits and trained models amplify the difference with every replayed token. Compare only the first replayed token of each seq, then restore

发布时间：2026-10-27 16:18

抓取时间：2026-10-27 16:38

来源机构：ggml-org