Skip to main content

JAX

Open-source high-performance numerical computing and automatic differentiation library for machine learning, scientific computing and optimisation.

Main category
Scientific computing
Cost
See the official site for cost
Licence type
Open source
Platforms
Linux / macOS / Windows
Visit JAX Code repository

Version-watched: add it to favourites and confirmed new releases will show on your favourites page (stored in this browser only).

Details

JAX is an open-source Python numerical computing library supporting automatic differentiation, JIT compilation, GPU/TPU acceleration and a NumPy-style interface. It suits machine learning research, scientific computing, optimisation problems, numerical simulation and high-performance computing experiments.

Use cases: scientific computing, automatic differentiation, machine learning, optimisation, high-performance computing

Note: JAX has a steeper learning curve than plain NumPy and suits users with Python and numerical computing experience.