jaxtyping VS diffrax

You want runtime typechecking.

See either beartype [1] or typeguard [2]. And if you're doing any kind of array-based programming (JAX or not), then jaxtyping [3].

[1] https://github.com/beartype/beartype/

[2] https://github.com/agronholm/typeguard

[3] https://github.com/patrick-kidger/jaxtyping

Try using [jaxtyping](https://github.com/google/jaxtyping).

It also supports numpy/pytorch/etc.

Since you mention ML use-cases, you might like jaxtyping.

jaxtyping: rich shape & dtype annotations for arrays and tensors (also supports PyTorch/TensorFlow/NumPy);

Heads-up that my newer jaxtyping project now exists.

As other folks have commented, type hints are now a big deal. For static typing the best checker is pyright. For runtime checking there is typeguard and beartype. These can be integrated with array libraries through jaxtyping. (Which also works for PyTorch/numpy/etc., despite the name.)

Did you use https://github.com/patrick-kidger/diffrax ?

Compute the Christoffel symbols for a metric and solve the geodesic equation to find the EOM using automatic differentation powered differential equation solvers, such as diffrax

Optimistix has high-level APIs for minimisation, least-squares, root-finding, and fixed-point iteration and was written to take care of these kinds of subroutines in Diffrax.

Diffrax (https://github.com/patrick-kidger/diffrax).

Here is the GitHub: https://github.com/patrick-kidger/optimistix

The elevator pitch is Optimistix is really fast, especially to compile. It

Sure. So I've got some PyTorch benchmarks here. The main take-away so far has been that for a neural ODE, the backward pass takes about 50% longer in PyTorch, and the forward (inference) pass takes an incredible 100x longer.

FWIW this worked for me. :D My full-time job is now writing JAX libraries at Google. Equinox for neural networks, Diffrax for differential equation solvers, etc.

It's relatively early days yet, but JAX is in the process of developing its nascent scientific computing / scientific machine learning ecosystem. Mostly because of its strong autodifferentiation capabilities, excellent JIT compiler etc. (E.g. to show off one of my own projects, Diffrax is the library of diffeq solvers for JAX.)

Diffrax - solving ODEs with Jax and computing it's derivatives automatically functools - love partial and lru_cache fastprogress - simpler progress bar than tqdm

At least prior to this announcement: JAX was much faster than PyTorch for differentiable physics. (Better JIT compiler; reduced Python-level overhead.)

E.g for numerical ODE simulation, I've found that Diffrax (https://github.com/patrick-kidger/diffrax) is ~100 times faster than torchdiffeq on the forward pass. The backward pass is much closer, and for this Diffrax is about 1.5 times faster.

It remains to be seen how PyTorch 2.0 will compare, or course!

Right now my job is actually building out the scientific computing ecosystem in JAX, so feel free to ping me with any other questions.