Does it support 16:9 and 9:16
Hi, I want to generate some images which resolutions are 1920x1080 and 1080x1920. I wonder if I can do that with cascade?
Just Try
Just Try
Thank you for your time, but of course I have tried that. I set the resolution to 720 by 1280, but the program throws an error. So, I am asking if there is a way to use this model to generate images in 16:9 and 19:6 aspect ratios.
RuntimeError: pixel_unshuffle expects height to be divisible by downscale_factor, but input.size(-2)=181 is not divisible by 2
Just Try
Thank you for your time, but of course I have tried that. I set the resolution to 720 by 1280, but the program throws an error. So, I am asking if there is a way to use this model to generate images in 16:9 and 19:6 aspect ratios.
RuntimeError: pixel_unshuffle expects height to be divisible by downscale_factor, but input.size(-2)=181 is not divisible by 2
I also encountered this problem. The solution is to set the length and height to a multiple of 256
Just Try
Thank you for your time, but of course I have tried that. I set the resolution to 720 by 1280, but the program throws an error. So, I am asking if there is a way to use this model to generate images in 16:9 and 19:6 aspect ratios.
RuntimeError: pixel_unshuffle expects height to be divisible by downscale_factor, but input.size(-2)=181 is not divisible by 2
I also encountered this problem. The solution is to set the length and height to a multiple of 256
Thanks! I love u!!!
The diffusers library contains a bug that restricts output to multiples of 256. The github repo code allows any multiple of 8.
The diffusers library contains a bug that restricts output to multiples of 256. The github repo code allows any multiple of 8.
Thank you! I'll check the github repo code.