b8947
📦 llama-cppView on GitHub →
✨ 1 features
Summary
This release updates the pre-compiled binaries for numerous platforms, including macOS, Linux, Android, Windows, and openEuler, and introduces a minor logic change in model download preference.
✨ New Features
- Model download logic now prefers the q8_0 quantization format if q4_k is unavailable.