NeurIO is an open source Python library for deploying deep neural networks on embedded (edge) devices. It a comprehensive and flexible tool for deploying neural networks on embedded (edge) devices, that lets researchers push the state-of-the-art in low-power machine learning and developers easily build and test neural networks for IoT.
NeurIO was originally developed by researchers and engineers working in the ANDANTE European Project, with the goal of democratizing AI accelerators for edge devices.
Note: Only Python 3.9 is currently supported.
To use NeurIO, clone the current release:
$ git clone https://github.com/csem/neurio
Then, install the dependencies:
$ pip install -r requirements.txt
To build the documentation, you need to install the following dependencies:
$ pip install sphinx sphinx_book_theme
Then, you can build the documentation by running the following commands:
$ cd neurio/docs/
$ sh build_doc.sh
The documentation contains all the information for getting started using NeurIO.
You can then open the documentation by opening docs/_build/html/index.html
in your browser.
If you want to contribute to NeurIO, be sure to review the contribution guidelines. This project adheres to NeurIO's code of conduct. By participating, you are expected to uphold this code.
We use GitHub issues for tracking requests and bugs. The NeurIO project strives to abide by generally accepted best practices in open-source software development.
GNU General Public License (GPL), version 3 : GPLv3