ggmlv0.24.0
- Updated
Tensor library for machine learning
- Project website: https://github.com/ggml-org/ggml
#Features
No default features.
#blas
Enable BLAS support
Host Dependencies:
2 transitive dependencies:
- vcpkg-cmake (through openblas)
- vcpkg-cmake-config (through openblas)
#cuda
Enable CUDA support
Platforms: !(windows & staticcrt)
- ✅linux
- ✅osx
- ✅windows
Host Dependencies:
No dependencies.
#metal
Enable Metal support
Platforms: osx
- ✅osx
- ❌
linux - ❌
windows
Dependencies:
No dependencies.
Host Dependencies:
No dependencies.
#opencl
Enable OpenCL support
Platforms: !arm32
- ✅linux
- ✅osx
- ✅windows
Host Dependencies:
2 transitive dependencies:
- vcpkg-cmake (through opencl)
- vcpkg-cmake-config (through opencl)
#openmp
Enable OpenMP support
Platforms: !osx
- ✅linux
- ✅windows
- ❌
osx
Dependencies:
No dependencies.
Host Dependencies:
No dependencies.
#vulkan
Enable Vulkan support
Dependencies:
and 2 transitive dependencies:
- vulkan-headers (through vulkan)
vulkan-loader (through vulkan)
Platforms:
!android
Host Dependencies:
and 3 transitive dependencies:
- spirv-tools (through shaderc)
- vcpkg-cmake (through glslang)
- vcpkg-cmake-config (through glslang)
#Dependencies
No dependencies.
#Used By
- hanjingo-high-jump (feature
cuda) - llama-cpp
- whisper-cpp
- whisper-cpp (feature
cuda) - whisper-cpp (feature
metal) - whisper-cpp (feature
vulkan)
#Contributors
Stefano Sinigardi
Kai Pastor
talregev
Ankur Verma
Chris Sarbora
Billy O'Neal
miyan
Weihang Ding
Reeshabh Choudhary