Metadata-Version: 2.4
Name: perforated
Version: 0.1.0
Summary: Perforated AI - parameter-efficient neural network training by Perforated AI
Project-URL: Homepage, https://perforated.ai
Project-URL: Repository, https://github.com/PerforatedAI/PerforatedAI
Author-email: Perforated AI <contact@perforated.ai>
License: Proprietary
Keywords: backpropagation,deep learning,neural networks,parameter efficiency,pytorch
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.8
Description-Content-Type: text/markdown

# Perforated

**Perforated Backpropagation** by [Perforated AI](https://perforated.ai)

A parameter-efficient neural network training algorithm that adds dendritic computation layers to standard neurons, enabling models to achieve greater capability density.

## Status

This package is a placeholder. Full SDK coming soon.

## Links

- [GitHub](https://github.com/PerforatedAI/PerforatedBackpropagation)
- [ArXiv Paper](https://arxiv.org/abs/2501.18018)
- [Patent US11580403B2](https://patents.google.com/patent/US11580403B2)
