nexport 0.4.4

A Python package for exporting the weights and biases of neural networks.

---

Overview

nexport is a lightweight Python 3.10+ package which enables neural network developers to export the weights and biases of trained networks to useful file types.

Table of contents

• Overview
• Table of contents
• Current support
• Install & use
• Objectives
• History
• Credits
• License
• Links

Current support

Filetype	PyTorch	Keras/TensorFlow
Export	Import	Export	Import
Text (.txt)	✅	✅	🚧	❌
JSON (.json)	✅	🚧	❌	❌
CSV (.csv)	❌	❌	❌	❌
XML (.xml)	❌	❌	❌	❌

Install & use

1. From terminal: pip install nexport
2. From python environment: import nexport

Objectives

• Export weights and biases to human-readable file
• Ensure compatability with all popular neural network development software

History

This package is intended to be used in conjunction with inference-engine. As such, nexport was developed by the inference-engine developers to enable compatability between the two softwares. nexport does this by exporting the weights and biases from networks compiled in PyTorch, Keras, and TensorFlow into standardized human-readable files. These files can be read by inference-engine to instantiate the netwoks in Fortran 2018 for inference.

Credits

nexport was created and is currently maintained by Jordan Welsman. Parts of this project were based on prior work by Tan Nguyen.

License

nexport is developed and distributed under a Berkeley Laboratory modified BSD license.

Note See LICENSE for more details.

Links

:file_folder: See this project on GitHub

:gift: See this project on PyPI

:cat: Follow me on GitHub

:briefcase: Connect with me on Linkedin

:email: Send me an email

:thought_balloon: Based on this project