Florian

Sure, I'm fine with anything :)

Hey, point clouds are currently not supported. It could probably be done in pure Python, although you would have to figure out a way to handle occlusion properly (so that...

One way of reverting back to a point cloud would be to extract the vertices of the triangles (although this wouldn't necessarily coincide with the original point cloud used to...

It's hard to say what a segmentation fault could be caused by without more information. You could try adding some print statements in the C++ function where the error happens...

It looks like the traceback of the error is in a codepath that can only execute if `"jax" in sys.modules`. So likely there is a module with name "jax" that...

The problem should also be fixed with the latest source version of einx (see [this commit](https://github.com/fferflo/einx/commit/affbd5c94064e344f52fb0fdb5e34d3b8e7bc686)), so you could try updating einx: ``` pip install --upgrade git+https://github.com/fferflo/einx.git ```

Hi, thanks for bringing this up! Support for mx.einsum (einx.dot) is already implemented, although not in a release version yet: https://github.com/fferflo/einx/commit/0ad7af3e675f6cd611cd0e38a7716f4eec81f66a It will be included in the next release of...

``einx.param`` requires different arguments based on the backend that you are using. In Torch, you always have to define model parameters in the constructor of a module (as ``nn.parameter.UninitializedParameter`` if...

This works on my end: ``` import jax.numpy as jnp import einx b = 4 h = 80 w = 128 c = 3 i = 10 image = jnp.zeros((b,...

The specific return type depends on the value of the string expression (i.e. if there is a comma on the right side or not), so I am not sure if...