llama.cpp b10338 修复 MoE 模型保存加载 bug

AI 摘要

llama.cpp 发布 b10338 版本，修复了模型保存时共享专家 FFN 长度键被覆盖的 bug，该问题导致 MoE 模型（如 qwen2moe、granite-moe 等）在保存后无法重新加载。修复后，共享专家和分块 FFN 长度均能正确保存，并添加了相应测试。

正文节选

<details open> model-saver : fix expert shared/chunk FFN length key clobber (#26693) The saver called add_kv with LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH twice, the second time passing n_ff_chexp. gguf_set_val_u32 removes-then-appends, so the second call clobbers the first: the saved shared_feed_forward_length ends up as n_ff_chexp (0 for every arch except GroveMoE), and expert_chunk_feed_forward_length is never written at all. So a save->load roundtrip of any MoE model with a shared expert

发布时间：2026-08-10 18:51

抓取时间：2026-08-10 18:56

来源机构：ggml-org