Justin Kerr
Justin Kerr
A quick idea would be changing `eps` in that function to something larger (maybe 1e-4 or something), it's possible lines 373 and 374 are exploding when the denominator is too...
Tested to 30k iterations with eps=1e-3 and it seems to work
Would be good to do a quality comparison to make sure it doesn't affect results
Interesting, I've never trained with a dataset that large. The reason that check is there is because if a camera isn't observed before the gaussians get culled, their opacity won't...
@brentyi This seems related to the depth rendering scaling factor
Have you looked at the trained nerf in the viewer? how does the actual model itself look?
The `outputs['depth']` from the model is a real float depth value, the viewer just converts it to a colormap. If you want the raw values you can save it in...
Potentially there's an issue with camera downscaling? the train dataset downscales cameras automatically and maybe the resolution of the image you're visualizing doesn't match these downscaled intrinsics?
I believe that focal length in nerfstudio is defined in terms of pixel units, so you might also need to rescale the focal length to match the new height/width. You...
I've never used Windows so I'm not entirely sure, maybe if you check the line of code the error is being thrown from you might be able to tell what's...