Pinned Loading

1. llama.cpp-triattention-mtp llama.cpp-triattention-mtp Public

   llama.cpp-triattention-mtpllama.cpp fork: Qwen3 MTP (PR #22673) + TriAttention for long-context inference on consumer GPUs上游 MTP (#22673) + TriAttention 合并构建，面向消费级 GPU 的长上下文推理

   C++ 1