keras-custom 0.1.1

Custom Layers in keras 3

Explore KerasCustom docs »

👋 Welcome to keras custom documentation!

Keras Custom is a Python library that extends Keras with custom, non-native classes and modules designed to enhance model manipulation. This library introduces powerful new features, including advanced model analysis tools and utilities, that are not available in Keras by default. It provides a clear, modular framework built on top of Keras, making it an invaluable tool for researchers, educators, and developers working in deep learning.

The new non-native classes in Keras Custom enable users to efficiently analyze, modify, and optimize Keras-based neural models for a variety of downstream tasks. These custom components open up new possibilities for customizing and extending Keras models beyond the built-in functionality.

Whether you're exploring new architectures, conducting research, or building complex deep learning workflows, Keras Custom offers the flexibility and power to streamline your work.

📚 Table of contents

• 📚 Table of contents
• 🔥 Tutorials
• 🚀 Quick Start
• 📦 What's Included
• 👍 Contributing
• 🙏 Acknowledgments
• 📝 License

🚀 Quick Start

You can install keras custom directly from pypi:

pip install keras_custom

In order to use keras custom, you also need a valid Keras installation. keras custom supports Keras versions 3.x.

🔥 Tutorials

Tutorial Name	Notebook
Model splitting - Splitting an existing models into a sequence of nested models
Model switching - Conversion to channel first to channel last and vice versa
Model fusion - Combining a sequence of models into a single model with only Layers	Stay tuned !

Documentation is available online.

👍 Contributing

#To contribute, you can open an #issue, or fork this #repository and then submit changes through a #pull-request. We use black to format the code and follow PEP-8 convention. To check that your code will pass the lint-checks, you can run:

tox -e py39-lint

You need tox in order to run this. You can install it via pip:

pip install tox

🙏 Acknowledgments

This project received funding from the French program within the Artificial and Natural Intelligence Toulouse Institute (ANITI). The authors gratefully acknowledge the support of the DEEL project.

📝 License

The package is released under MIT license.