-
equinox
Elegant easy-to-use neural networks + scientific computing in JAX. https://docs.kidger.site/equinox/
-
diffrax
Numerical differential equation solvers in JAX. Autodifferentiable and GPU-capable. https://docs.kidger.site/diffrax/
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
A neural network library for JAX: Equinox. IMO the simplest/most-generic one out there.
Whilst this has been getting a bit of traction in the community, this started out life as a personal project. I needed a way to describe generic parameterised functions (not just neural networks) to be able to describe differential equation solvers, and found that existing options (Flax/Haiku) were overly-tailored to neural networks, and couldn't handle my use case.
Related posts
-
Show HN: Optimistix: Nonlinear Optimisation in Jax+Equinox
-
PyTorch 2.0
-
[D] Adjoint Sensitivity Method vs Reverse Mode Autodiff
-
Solving system of coupled differential equations using Runge-Kutta in python
-
Why is Python used by lots of scientists to simulate and calculate things, although it is pretty slow in comparison to other languages?