About figure11 in the paper

[lihuining]

Hello, author. I would like to ask about Figure 11 in the paper. Is the image with 64 perspectives generated all at once? When I was replicating the paper, I attempted to generate the perspectives separately instead of all at once, but this resulted in different subjects each time. However, generating so many images at once seems to exceed the available GPU memory. How was this process handled in the paper?