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
In what sense are you trying to change the size of the output? If you have a pre-trained model with another output size, you can use the --size argument to tell the model to expect this alternative size. If you're asking whether a pre-trained 1024x1024 model can be used to produce images of other resolutions - the answer is no. You'll have to resize the output yourself. For decreasing resolution you can probably just work with PIL's resize. If you want to increase resolution, I'd have a look at super resolution models like ESRGAN.
I looked through the closed issues and found some tips about the quality of images, increasing iterations, and decreasing learning_rate. Also, tried to use the ESRGAN model, and the resolution is increased but the quality and blur effect remained the same.
Right now the size of the output is 1024x1024 pixels. Can I change the size of the output?
The text was updated successfully, but these errors were encountered: