There are differences in the output image between the recently added inpaint function and the previous generation function, other than background compositing. Specifically, the previous generation function preserved the aspect ratio of the composite object, whereas the inpaint function fills the background image and the aspect ratio appears to be out of sync. What is the difference?
Is it because the model has changed? Or is it because the post- or pre-processing has changed?
There are differences in the output image between the recently added inpaint function and the previous generation function, other than background compositing. Specifically, the previous generation function preserved the aspect ratio of the composite object, whereas the inpaint function fills the background image and the aspect ratio appears to be out of sync. What is the difference?
Is it because the model has changed? Or is it because the post- or pre-processing has changed?