LLaMA.cpp

The ggerganov/llama.cpp repository on GitHub contains a port of Facebook's LLaMA model in C/C++, as well as the high-performance, dependency-free Whisper automatic speech recognition (ASR) model. Whisper is optimized for Apple silicon and supports several architectures, including low memory usage and integer quantization. It can run on various platforms, including the browser, and is lightweight and portable for easy integration into different applications.