tensorflow/lucid
A TensorFlow library for visualizing and understanding neural network internals through techniques like DeepDream and feature visualization.
Not currently ranked — collecting fresh signals.
star history
Lucid provides infrastructure and tools for neural network interpretability research, enabling researchers to visualize what neural networks learn and how they process information. It offers Jupyter notebook-based tutorials and implements visualization techniques like DeepDream, activation maximization, and feature inversion. The library is built on TensorFlow 1.x and is designed specifically for research exploration rather than production use.
Frequently asked
- What is tensorflow/lucid?
- A TensorFlow library for visualizing and understanding neural network internals through techniques like DeepDream and feature visualization.
- Is lucid open source?
- Yes — tensorflow/lucid is open source, released under the Apache-2.0 license.
- What language is lucid written in?
- tensorflow/lucid is primarily written in Jupyter Notebook.
- How popular is lucid?
- tensorflow/lucid has 4.7k stars on GitHub.
- Where can I find lucid?
- tensorflow/lucid is on GitHub at https://github.com/tensorflow/lucid.