NeuraViz is a web application for visualizing the underlying structure of neural networks, primarily for teaching purposes.
Explore the docs »
Report Bug
·
Request Feature
Neural networks are notoriously difficult to visualize for students and experienced engineers alike due to their scale and complexity. Too combat this problem, NeuraViz was devised to provide a simple, intuitive interface for visualizing the structure of neural networks. While it can be used on larger networks, such as those used for real-world problems, visualizing networks of that scale has diminishing returns. The structure of individual nodes and links at a scale of even hundreds of nodes just isn't meaningful. Instead, NeuraViz is intended to be used for smaller networks, such as those one might use to learn about how neural networks work. For example, the image above shows the visualization of a network trained to classify the Iris dataset, a famous toy example used primarily for teaching about neural networks.
This project is written in Python with Quart and Typescript with Svelte and FlowBite Svelte. In addition, the following libraries are used:
Each of their respective licenses apply to their binaries and their use in this project. Their licenses can be found at the links above.
Contributions are what make the open source community such an amazing place to be learn, inspire, and create. Any contributions you make are greatly appreciated.
git checkout -b feature/AmazingFeature
)git commit -m 'Add some AmazingFeature'
)git push origin feature/AmazingFeature
)If you find an issue in existing code, feel free to use the above procedure to generate a change, or open an issue for me to fix it.
Distributed under the BSD 3-Clause License. See LICENSE
for more information.
Bennett Wendorf - Website - bennettwendorf@gmail.com
Project Link: https://neuraviz.bennettwendorf.dev/