Metadata-Version: 2.4
Name: strideweave
Version: 0.0.0
Summary: A deep learning framework built on hierarchical layout algebra. Under active development; this release reserves the name.
Project-URL: Homepage, https://github.com/patrickk1998/strideweave
Project-URL: Repository, https://github.com/patrickk1998/strideweave
Project-URL: Issues, https://github.com/patrickk1998/strideweave/issues
Author-email: Patrick Krusiec <patrick.krusiec@gmail.com>
License-Expression: Apache-2.0
License-File: LICENSE
Keywords: cuda,deep-learning,distributed,gpu,layout,machine-learning,tensor
Classifier: Development Status :: 1 - Planning
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Typing :: Typed
Requires-Python: >=3.12
Description-Content-Type: text/markdown

# strideweave

A deep learning framework built on hierarchical layout algebra.

> **Status: placeholder.** This release reserves the name on PyPI. There is no
> public API yet. The project is under active development — follow the
> repository for progress.

## What it will be

`strideweave` treats tensors as data-class objects and builds on CuTe-style
hierarchical layouts as the unifying primitive. Sharding and replication fall
out as special cases of layout rather than a separate bolted-on type, and
orthogonal tensor properties compose through combinators instead of inheritance.

## Installation

Nothing to install yet.

```bash
pip install strideweave
```

## License

Apache-2.0. See [LICENSE](LICENSE).
