Stefan Baumann

Okay, admittedly I don't know much about creating pip packages. Looking back more closely at the terminal output I got when installing it, it seems like it's actually included, just...

In this code, gamma is hardcoded to depend on sigma_data, with gamma being chosen as gamma = sigma_data^-2. This, combined with the preconditioner compensation, leads to the formula you're seeing.

Thanks for the quick response! Regarding the question about `tcnn`, I was primarily interested as to why you seem to always use the `tcnn` version directly instead of your own...

That would be awesome if you could do an integration! If you need any help or guidance, let me know, and I'll happily help!

While I don't know that much about diffuser's design philosophy (otherwise, I would have created a PR myself already), I'm happy to answer any questions about the method and the...

Good catch! Yeah, we should consider changing this, although at this point (given that it also seems to work well with the double norm), we also have to consider whether...

It turns out that this is primarily a problem with the scaling: the scaling you need to get the the Blender/NeRF-Synthetic dataset running is different from the one you use....

Okay yeah, I guess that essentially ends up doing something very similar. The motivation for the factor of 4 (or rather $\sqrt{4^2 + 0.5^2}$) I mentioned was that this is...

Thank you very much! I will give it a try asap :)

I have a quick question about your implementation of the sparsity loss: you regularize the SDF values from the points where you already sample due to the main rendering. The...