cppgrad/Roadmap
Roadmap
Where cppgrad is headed — broader op coverage, custom backends, and a deeper test suite.
cppgrad is alpha and evolving. This page captures the direction; for the
live state, follow the repository.
In progress
- Expanding the test suite. Coverage across tensor manipulation, autograd correctness, and edge cases is a work in progress.
- More operations and gradient functions. Growing the catalog of differentiable ops beyond the current set.
Planned / exploratory
- Custom backends. A cleaner path to swap ArrayFire for hand-written CUDA or OpenCL backends. See Backends.
- Higher-level neural network building blocks layered on top of the tensor and autograd core.
Contributing
Contributions are welcome:
- Fork the repo and create a feature branch.
- Write tests and documentation alongside your change.
- Open a PR.
Community & support
- Overview: deepwiki.com/Research-Commons/cppgrad
- Issues: github.com/Research-Commons/cppgrad/issues
- Discussions: github.com/Research-Commons/cppgrad/discussions
Licensed under Apache 2.0.