-
Notifications
You must be signed in to change notification settings - Fork 1.2k
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.
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
Fix opencv 4.5.2 compatibility #635
Conversation
BottomUpRandomAffine fails with opencv-python>=4.5.2 because a \ single valued numpy.ndarray is no longer a valid data type for \ image_size in cv2.warpAffine. This commit fix it by repacing \ image_size with image_size.item().
@@ -1,4 +1,5 @@ | |||
albumentations>=0.3.2 | |||
loguru |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@zengwang430521 Do you think we should add loguru into requirements to support smplx>=0.1.27, or just limit smplx==0.1.26?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let's update smplx to 0.1.27. I think this updating will not change the performance.
Codecov Report
@@ Coverage Diff @@
## master #635 +/- ##
==========================================
- Coverage 82.51% 82.35% -0.16%
==========================================
Files 167 167
Lines 11992 11872 -120
Branches 1927 1927
==========================================
- Hits 9895 9777 -118
+ Misses 1620 1618 -2
Partials 477 477
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
Is this a bug of smplx? Please make a fix there |
Yeah, I submitted an issue there. We still need a work-around in mmpose before it's fixed in smplx. |
Maybe just use smplx==0.1.26 for safety? We haven't tested our mesh model with smplx==0.1.27 anyway. |
Adding loguru to requirements is better than pining smplx version. If you pin it now, then it is hard to unpin it later. |
I think it's okay to update the smpl version and it will not change the performance. |
* fix opencv-python 4.5.2 compatibility BottomUpRandomAffine fails with opencv-python>=4.5.2 because a single valued numpy.ndarray is no longer a valid data type for image_size in cv2.warpAffine. This commit fix it by repacing image_size with image_size.item(). * fix smplx 0.1.27 compatibility smplx>=0.1.27 requires loguru, which will not be automatically installed when installing smplx. So we add loguru into mmpose requirements.
* fix opencv-python 4.5.2 compatibility BottomUpRandomAffine fails with opencv-python>=4.5.2 because a single valued numpy.ndarray is no longer a valid data type for image_size in cv2.warpAffine. This commit fix it by repacing image_size with image_size.item(). * fix smplx 0.1.27 compatibility smplx>=0.1.27 requires loguru, which will not be automatically installed when installing smplx. So we add loguru into mmpose requirements.
* fix opencv-python 4.5.2 compatibility BottomUpRandomAffine fails with opencv-python>=4.5.2 because a single valued numpy.ndarray is no longer a valid data type for image_size in cv2.warpAffine. This commit fix it by repacing image_size with image_size.item(). * fix smplx 0.1.27 compatibility smplx>=0.1.27 requires loguru, which will not be automatically installed when installing smplx. So we add loguru into mmpose requirements.
opencv-python 4.5.2 compatibility
BottomUpRandomAffine fails with opencv-python>=4.5.2 because a
single valued numpy.ndarray is no longer a valid data type for
image_size in cv2.warpAffine. This commit fix it by repacing
image_size with image_size.item().
smplx 0.1.27 compatibility
smplx>=0.1.27 requires loguru, which will not be automatically installed
when installing smplx. So we add loguru into mmpose requirements.