b8424
📦 llama-cppView on GitHub →
✨ 1 features
Summary
This release introduces an optimization for Vulkan by enabling 4-at-a-time dequantization for the iq4_xs format. It also provides numerous pre-compiled binaries for various operating systems and hardware configurations.
✨ New Features
- Implemented dequantization of iq4_xs format 4 values at a time for Vulkan backend.