This repo contains a Nix package that can be used to build custom machine learning kernels for PyTorch. The kernels are built using the PyTorch C++ Frontend and can be loaded from the Hub with the kernels Python package.
This builder is a core component of the larger kernel build/distribution system.
We provide a Docker image with which you can build a kernel:
# navigate to the activation directory
cd examples/activation
# then run the following command to build the kernel
docker run --rm \
-v $(pwd):/kernelcode \
ghcr.io/huggingface/kernel-builder:latestThis will build the kernel and save the output in the build directory in
the activation folder.
- Writing Hub kernels
- Building kernels with Docker
- Building kernels with Nix
- Local kernel development (IDE integration)
The generated CMake build files are based on the vLLM build infrastructure.