I noticed the generated images are kind of blurry, and not as sharp as the images generated by SD.
For example, from the figure shown in the paper:
The trees and the train from SD2 are much sharper compared with the images on the last column.
I also tested the image variations generation using the shared fine-tuned SD model, it looks like the generated images are blurry too (please zoom in):
There are even some jpeg effects if you zoom in.
Is it expected? How can I solve it? Thank you very much!
Hi,
I noticed the generated images are kind of blurry, and not as sharp as the images generated by SD. For example, from the figure shown in the paper:
The trees and the train from SD2 are much sharper compared with the images on the last column.
I also tested the image variations generation using the shared fine-tuned SD model, it looks like the generated images are blurry too (please zoom in):
There are even some jpeg effects if you zoom in.
Is it expected? How can I solve it? Thank you very much!