Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix a bug when the device of manolayer is cpu and use_pca=False #26

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Fix a bug when the device of manolayer is cpu and use_pca=False #26

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
7 changes: 4 additions & 3 deletions manopth/rotproj.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,14 +8,15 @@ def batch_rotprojs(batches_rotmats):
for rot_idx, rotmat in enumerate(batch_rotmats):
# GPU implementation of svd is VERY slow
# ~ 2 10^-3 per hit vs 5 10^-5 on cpu
U, S, V = rotmat.cpu().svd()
rotmat = torch.matmul(U, V.transpose(0, 1))
_device = rotmat.device
U, S, V_T = torch.linalg.svd(rotmat.cpu())
rotmat = torch.matmul(U, V_T)
orth_det = rotmat.det()
# Remove reflection
if orth_det < 0:
rotmat[:, 2] = -1 * rotmat[:, 2]

rotmat = rotmat.cuda()
rotmat = rotmat.to(_device)
proj_batch_rotmats.append(rotmat)
proj_rotmats.append(torch.stack(proj_batch_rotmats))
return torch.stack(proj_rotmats)