Silvan Weder

I assume you don't have ground-truth. You can just initialize a zero ground-truth grid (as you don't train on it) and process the reconstructed meshes/grids.

You can use the exact same [script](https://github.com/weders/RoutedFusion/blob/master/test_fusion.py). You just need to write a dataset class (similar to [this](https://github.com/weders/RoutedFusion/blob/master/dataset/shapenet.py)) with a get_grid(self, scene, truncation) method and a scenes property. The input...

You have to adjust the `setup.py` file such that it does not use `torch.cuda.is_available()` as the GPU is not exposed to the docker during `docker build .` The `force_cuda` flag...

`torch.cuda.is_available()` returns `True` if you run it on your machine? Something I also realized is that the CUDA versions have to perfectly match between PyTorch installation and CUDA Toolkit.

Thanks a lot for the answer! Is this code for the COLMAP conversion available?

The dashboard approach solved it for me on a Mac: https://devops.stackexchange.com/questions/15602/mac-m1-docker-desktop-no-longer-launches I.e, do the following steps: 1. Open the docker dashboard 2. Go to settings / advanced 3. Check the...

Hi @ericyangchen, From just looking at the results you provided, it seems like there is a channel flipping issue. The colors are not correct. Can you please provide me with...