You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I noticed that in the apply_se3_aug_con function for rotation augmentation, there are two options, ver=1 and ver=2. However, I observed that at line 364 ( ver=1) and line 368 ( ver=2) in augmentation.py, the rotation is handled differently. Specifically, in ver=1, the rotation matrix is applied via right multiplication, whereas in ver=2, it's applied via the transpose of the rotation matrix using left multiplication. Generally, these two approaches yield different results. I noticed that PerAct uses the former, while RVT uses the latter. Could you please explain why you are using the latter?
The text was updated successfully, but these errors were encountered:
Thank you for your great work.
I noticed that in the
apply_se3_aug_confunction for rotation augmentation, there are two options,ver=1andver=2. However, I observed that at line 364 (ver=1) and line 368 (ver=2) in augmentation.py, the rotation is handled differently. Specifically, inver=1, the rotation matrix is applied via right multiplication, whereas inver=2, it's applied via the transpose of the rotation matrix using left multiplication. Generally, these two approaches yield different results. I noticed that PerAct uses the former, while RVT uses the latter. Could you please explain why you are using the latter?The text was updated successfully, but these errors were encountered: