Open
Description
Having tested PSHuman with some samples, I find some problems that can probably be solved with SMPLX conditioning. According to Table 6 in the paper, SMPLX conditioning is very important for model performance.
Correct me if I'm wrong, but I think the currently released SMPL-Free version and thee SMPLX-based version have different weights. Therefore, I'd like to ask whether the SMPLX-based version of PSHuman will be released?
Thank you again for your great work~
Metadata
Metadata
Assignees
Labels
No labels