Open Duemellon opened 5 months ago
Don’t know if there is a gap between the implementation and the paper. Needs to be jointly debugged with algorithm engineers. We'll look at this next week.
Any news on this? I have the same problem.
@budui Take a look at this
Regardless of the model I use, at a certain point of complexity the results turn into a cartoony 3D render-quality. Whether it's a person, scenery, or otherwise.
Anyone else experiencing this?