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
When I use the demo demo on huggingface for image segmentation, I get great results:
But when I use the published pre-trained model for segmentation, I get poor results:
Do you know what the problem is? What can I do to reproduce the results of the demo?
The text was updated successfully, but these errors were encountered:
When I use the demo demo on huggingface for image segmentation, I get great results:


But when I use the published pre-trained model for segmentation, I get poor results:
Do you know what the problem is? What can I do to reproduce the results of the demo?
The text was updated successfully, but these errors were encountered: