Metadata-Version: 2.4
Name: matnets
Version: 3.2.4
Summary: A small experimental neural network library where neurons are represented as matrices.
License-Expression: MIT
License-File: LICENSE
Keywords: neural-networks,matrix,machine-learning,jax,non-classical-networks
Author: dsainvg
Requires-Python: >=3.11
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Provides-Extra: dev
Provides-Extra: docs
Requires-Dist: jax (>=0.4)
Requires-Dist: mkdocs (>=1.6) ; extra == "docs"
Requires-Dist: mkdocs-material (>=9.5) ; extra == "docs"
Requires-Dist: mypy (>=1.10) ; extra == "dev"
Requires-Dist: pytest (>=8.0) ; extra == "dev"
Requires-Dist: ruff (>=0.5) ; extra == "dev"
Project-URL: Documentation, https://dev.dsainvg.me/MATNETS/api
Project-URL: Homepage, https://dev.dsainvg.me/MATNETS
Project-URL: Issues, https://github.com/dsainvg/MATNETS/issues
Project-URL: Repository, https://github.com/dsainvg/MATNETS/
Description-Content-Type: text/markdown

# MATNETS

MATNETS is a JAX library for matrix-neuron neural network experiments.

In a traditional neural network, a neuron carries one scalar activation. In
MATNETS, each neuron carries an `n x n` matrix. A layer maps a stack of input
matrix-neurons to a stack of output matrix-neurons.

## Core Shape Contract

The core dense primitive uses:

```text
params.W: (q, p, n, n)
params.B: (q, n, n)
x:        (p, n, n)
output:   (q, n, n)
```

`p` is the input neuron count. `q` is the output neuron count. `n` is the matrix
size for every neuron.

## Read Next

- [Getting Started](getting-started.md): install MATNETS and run a first dense layer.
- [Concepts](concepts.md): understand matrix-neuron shapes and JAX transforms.
- [API Guide](api.md): see each public function and its expected shapes.
- [Examples](examples.md): run the included examples.
- [Development](development.md): run tests and local checks.

