It claims to be diffusion-based, but the main 2 differences from an approach like Stable-Diffusion is that (1) they only consider a single step, instead of a traditional 1000 and (2) they directly predict the value z^y instead of a noise direction. According to their analyses, both of these differences help in the studied tasks. However, isn't that how supervised learning has always worked ? Aside from having a larger model, this isn't very different from "traditional" depth estimation that don't claim anything to do with diffusion.
It also claims to have zero-shot abilities, but they fine-tune the denoising model f_theta on a concatenation of the latent image and apply a loss using the latent label. So their evaluation dataset may be out-of-distribution, but I don't understand how that's zero-shot. Asking ChatGPT to output a depth estimation of a given image would be zero-shot because it hasn't been trained to do that (to my knowledge).
I've recently had to implement a few kernels to lower the memory footprint and runtime of some pytorch function : it's been really nice because numba kernels have type hints support (as opposed to raw cupy kernels).
Which software do you use as an emergency backup right before doing a risky operation (such as an upgrade) ?I am going through the pain of restoring my files after a bad error on my part : I used déjà-dup (the default with Ubuntu 20.04.4) to back up the default home folder.Unfortunately, my update broke everything and I lost access to my session. I hoped that deja-dup would allow me to restore...
As a student, I have a desktop at one of my parents' house that I can control over ssh, this kind of features make remote development much easier and is often needed when I run an intensive task for hours. The experience with VSCode over ssh is really great. Some have pointed out local VMs, which is another use for this.