qwen.cpp and qwen2.cpp

These are ecosystem siblings serving different implementation needs: QwenLM/qwen.cpp is the official C++ inference engine for Qwen models, while yvonwin/qwen2.cpp is a community fork that extends the approach to support both Qwen2 and Llama3 models with potentially different optimizations.