llama.cpp and ggml reach v1.0, with semantic versioning and a macOS desktop app
The ggml-org team at Hugging Face released v1.0.0 of llama.cpp (libllama) and ggml (libggml) on November 13, 3.5 years after the project began.
- From now on, a breaking change to the public C API or ABI in each library's
include/headers bumps the major version. Backward-compatible features, backend updates and model support bump the minor version, and bug fixes with no API change bump the patch version. The team says the aim is a stable interface for the downstream applications that depend on the libraries. - The team lists pre-release artifacts, Docker, Debian, Homebrew, conda-forge, MacPorts, Winget and an XCFramework for iOS and macOS as distribution channels.
- It says llama.cpp supports more than 150 model architecture variants and the mainstream speculative decoding methods, with CI running on self-hosted runners from NVIDIA, AMD, Qualcomm, Arm, IBM and Intel.
- The post points to llama.cpp desktop apps for macOS and Windows, from the team behind llama.cpp and Hugging Face.
Sources
- Llama.cpp hits v1.0 with a native app for macOSHacker News · 4 comments
Get news like this as a push.
AI Tracker alerts you when something changes for the tools you use.