b8996
📦 llama-cppView on GitHub →
✨ 1 features🐛 1 fixes🔧 1 symbols
Summary
This release addresses vectorized handling issues in the ggml-webgpu backend, specifically fixing mul-mat and adding a vectorized variant for mul-mat-id. It also provides updated binaries across macOS, Linux, Android, Windows, and openEuler.
✨ New Features
- Added vectorized variant to mul-mat-id operation in ggml-webgpu backend.
🐛 Bug Fixes
- Fixed vectorized handling in mul-mat operation within ggml-webgpu backend.