Jax2D is a 2D rigid-body physics engine written entirely in JAX and based off the Box2D engine. Unlike other JAX physics engines, Jax2D is dynamic with respect to scene configuration, allowing heterogeneous scenes to be parallelised with vmap .

2025年1月27日 · Jax2D is a 2D rigid-body physics engine written entirely in JAX and based off the Box2D engine. Unlike other JAX physics engines, Jax2D is dynamic with respect to scene configuration, allowing heterogeneous scenes to be parallelised with vmap .

2022年5月13日 · Do you want to interpolation on regular grid (i.e. same space along one axis, but each axis can have different space)? It seems that interp2d of https://github.com/adam-coogan/jaxinterp2d accepts x and y with different sizes, but CartesianGrid doesn't. Okay I see. xp and yp can have different size.

jax.scipy.signal.convolve2d# jax.scipy.signal. convolve2d (in1, in2, mode = 'full', boundary = 'fill', fillvalue = 0, precision = None) [source] # Convolution of two 2-dimensional arrays. JAX implementation of scipy.signal.convolve2d(). Parameters: in1 – left-hand input to the convolution. Must have in1.ndim == 2.

Basic one-dimensional convolution is implemented by {func} jax.numpy.convolve, which provides a JAX interface for {func} numpy.convolve. Here is a simple example of 1D smoothing implemented via...

jax.numpy.atleast_2d# jax.numpy. atleast_2d (* arys) [source] # Convert inputs to arrays with at least 2 dimensions. JAX implementation of numpy.atleast_2d(). Parameters: arguments. (zero or more arraylike) arys (ArrayLike) Returns: an array or list of …

JAX provides a number of interfaces to compute convolutions across data, including: For basic convolution operations, the jax.numpy and jax.scipy operations are usually sufficient. If you want to do more general batched multi-dimensional convolution, the jax.lax function is …

2025年2月8日 · A jax-based 2d physics library, mainly intended to use for reinforcement learning research. Apache LICENSE 2.0 holds unless otherwise noted. vec2d.py is copied from PyMunk with the license-header as-is. Download the file for your platform. If you're not sure which to choose, learn more about installing packages. Uploaded using Trusted Publishing?

Import JAXFit before importing JAX since we need JAXFit to set all the JAX computation to use 64 rather than 32 bit arrays. Now let's define a 2D Gaussian using jax.numpy. You can construct...

The 2D discrete wavelet transform for JAX. Motivation The motivation for jax-wavelets is to replace the patching and unpatching transforms in Vision Transformer with transforms whose basis vectors are smooth and overlap, without increasing the number of floating point values input to and output from the model.