patrick-kidger/equinox
A JAX-based neural network library offering easy-to-use PyTorch-like syntax for building models compatible with the broader JAX ecosystem.
Not currently ranked — collecting fresh signals.
star history
Equinox is a deep learning framework built on JAX that provides neural network building blocks using PyTorch-like class syntax. It offers filtered transformation APIs, PyTree manipulation utilities, and advanced features like runtime errors, while remaining fully compatible with standard JAX operations like JIT compilation and automatic differentiation. The library treats models as PyTrees, allowing them to pass smoothly across JAX transformation boundaries.
Frequently asked
- What is patrick-kidger/equinox?
- A JAX-based neural network library offering easy-to-use PyTorch-like syntax for building models compatible with the broader JAX ecosystem.
- Is equinox open source?
- Yes — patrick-kidger/equinox is open source, released under the Apache-2.0 license.
- What language is equinox written in?
- patrick-kidger/equinox is primarily written in Python.
- How popular is equinox?
- patrick-kidger/equinox has 2.9k stars on GitHub.
- Where can I find equinox?
- patrick-kidger/equinox is on GitHub at https://github.com/patrick-kidger/equinox.