Matthew Tancik

Yea, @kerrj has an implementation. We are still testing it, the results currently aren't conclusive. If it ends up working well we will merge it into the nerfacto model.

Here is a branch with robustNeRF implemented. https://github.com/nerfstudio-project/nerfstudio/tree/justin/robustnerf It seems to work, however it is much slower to converge. Performance also depends a bit on hyperparam choices.

In some network setups, the webrtc appears to be disconnected despite the connection being successful. If the viewer is working for you, you can safely ignore that status.

The mesh generation method is fairly simplistic at the moment so results like this aren't that unprecedented. If you want a better mesh, consider exporting at a point cloud and...

I'm not really sure what is going on. The colors are predicted by casting small around the surface of the object along the normals. These rays might have poor normals,...

> (Not sure what the readthedocs failure is from, the details link goes to 404) The error is `/home/docs/checkouts/readthedocs.org/user_builds/plenoptix-nerfstudio/checkouts/1390/docs/quickstart/installation.md:162: WARNING: 'myst' reference target not found: #pytorch`

You can run `python scripts/docs/build_docs.py --clean-cache` to see the error locally. If you add `(pytorch)=` at line 20 and remove the `#`, it should work.

Looks like a pytorch cuda error. Maybe try reinstalling pytorch and checking your cuda version. https://github.com/pytorch/pytorch/issues/20860

Looping in @zmurez. Multiple camera support is enabled on the nerfstudio side (https://github.com/nerfstudio-project/nerfstudio/pull/1651)

What kind of camera parameters, focal length, lens distortions? If so, we don't currently support that.