-
Notifications
You must be signed in to change notification settings - Fork 3
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 demos on other platforms #95
Conversation
I need to debug further but I got this error
|
It looks like a torch issue for me. Putting it here in case you have seen this before.
|
@edyoshikun looks like you're using a torch build that is not compatible with numpy v2. Probably downgrade numpy. Also check if your python is native arm64. |
Actually maybe upgrade torch to 2.3: pytorch/pytorch#107302 |
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.
Took a bit of fuzzing around on my side (my conda version was old...) so updating conda did the trick.
This worked well for the VSNeuromast and VSCyto3D on M1.
Thank you @ziw-liu
* load statistics explicitly to avoid autocasting to float64 this allows the GPU transfer on MPS see Lightning-AI/pytorch-lightning#16213 * disable multiprocssing by default to avoid fork/spawn difference * remove unused patch size argument * black
Fixed the issue of normalization statistics (
float
from JSON) being auto-casted totorch.float64
so that MPS works on macos_arm64.Also disabled multiprocessing. Or we can put the whole thing in a
if __name__ == '__main__':
guard.