Faceswap Forum

I have not yet understood Phaze-A but is Progressive Image resize training supported in Phaze-A?
I have read about it here : - https://towardsdatascience.com/explaine ... cb2be0f431

Is it possible that in Faceswap we can slowly increase image resolution while training on the same model?

I have looked into Progressive growth, but decided against adding it, as it impacts accuracy of the model, and adds complication (to some already very complicated) to the user experience.

I am happy to revisit if anyone can demonstrate that it really adds value, but I'm yet to be convinced.