karpathy/convnetjs

Health v2 · maintenance only

Adoption not part of the score

11197 stars · 2071 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

ConvNetJS is a JavaScript library for defining and training neural networks, including convolutional networks and an experimental Deep Q Learning module, entirely in the browser. It ships with interactive demos for classification, regression, autoencoders, and reinforcement learning.

Use cases

• train a neural network in the browser without a server
• learn how convolutional neural networks work interactively
• run MNIST or CIFAR-10 training demos client-side
• experiment with deep Q learning reinforcement learning in JavaScript
• teach deep learning concepts with visual 2D classification demos
• build a small neural network demo for a web page

When to choose

• you want client-side neural network training with no backend
• you need educational, visualizable deep learning demos
• you want a lightweight dependency-free JS neural net implementation

When to avoid

• you need production-grade performance or GPU acceleration
• you need actively maintained software or working npm installs
• you need modern architectures like transformers or large-scale training

Facets

library · maturity abandoned

deep-learning machine-learning reinforcement-learning deep-learning machine-learning browser-extensions browser neural-networks convolutional-networks javascript education demos no-longer-maintained nodejs

Member repositories

For agents

markdown · JSON · MCP: product_card(name="karpathy/convnetjs")