tencent-ailab/Frequency_Aug_VAE_MoESR

Why are the images generated after going through VAE different from the images generated directly from stable diffusion?

Eaven21 opened this issue · 2 comments

Eaven21 commented

Hello, I have run the code you provided in Txt2Img with FA_VAE, but I have noticed a discrepancy between the images generated through VAE and the images generated directly from stable diffusion.

cswry commented

Hello, I have run the code you provided in Txt2Img with FA_VAE, but I have noticed a discrepancy between the images generated through VAE and the images generated directly from stable diffusion.

Can you give some examples? Different seed would results in different results even if the same model and the same text prompt

amandaluof commented

Hello, I have run the code you provided in Txt2Img with FA_VAE, but I have noticed a discrepancy between the images generated through VAE and the images generated directly from stable diffusion.

It seems there is something different between the two runs, like random seed. You can compare their latents to check.