Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

read_obj returns a face list of size 1 x 3 * n_faces, instead of 3 x n_faces

See original GitHub issue

🐛 Bug

When using read_obj from torch_geometric.io to load an obj file, the face field of the returned data object is of size 1 x 3 * n_faces instead of 3 x n_faces. This was not yet the case in v. 1.4.3.

To Reproduce

Steps to reproduce the behavior:

Load obj file with torch_geometric.io.read_obj
print(data.face.size())
output: torch.Size([1, n_faces * 3]) (in this case [1, 1500])

Expected behavior

output: torch.Size([3, n_faces]). This is correct for read_ply.

Environment

OS: Linux (Ubuntu 20.04.1)
Python version: 3.7.4
PyTorch version: 1.6.0
CUDA/cuDNN version: 10.2
GCC version: 9.3.0
Any other relevant information: PyG version: 1.6.1

Additional context

I’m investigating the cause and will submit a pull request when I’ve fixed the issue.

Issue Analytics

State:
Created 3 years ago
Comments:5 (2 by maintainers)

Top GitHub Comments

1reaction

rusty1scommented, Aug 26, 2020

Thanks for this issue and your valuable bug tracing. This is very helpful! I fixed the code by using t.split("/")[0]. For later releases, we can think of using an external library to allow for faster loading and less code maintenance.

I’m closing this issue for now, feel free to re-open in case there are any other concerns.

0reactions

rusty1scommented, Sep 3, 2020

Ok, I reverted to the old version for now 😃