poloclub/ganlab

Health v2 · maintenance only

Adoption not part of the score

1534 stars · 419 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

GAN Lab is an interactive, browser-based visualization tool for learning and experimenting with Generative Adversarial Networks. Built with TensorFlow.js, it lets users interactively train GANs on 2D data distributions and visualize their inner workings entirely in the browser.

Use cases

• learn how GANs work interactively
• visualize GAN training dynamics in the browser
• experiment with generative adversarial networks on 2D data
• teach deep learning concepts to students
• play with GANs without installing anything
• understand generator and discriminator behavior visually

When to choose

• you want an interactive, zero-install way to learn GAN fundamentals
• you need a visual teaching aid for deep learning courses
• you want to experiment with GANs on simple 2D distributions in a browser

When to avoid

• you need to train GANs on real images or large datasets
• you require production-grade or GPU-cluster training
• you need a programmable deep learning framework rather than a visual demo

Facets

application · maturity maintenance

machine-learning deep-learning data-visualization deep-learning machine-learning data-visualization education browser gan tensorflowjs interactive-visualization in-browser-training learning-tool education web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="poloclub/ganlab")