when I use stable-diffusion-2-1(768ema), something wrong happened. It seems that 768ema can't correctly sample noise and inverse image. but 2-1-base works well. Is there any differences between them? How can I use stable-diffusion-2-1(768ema)? the cause I studying is unit4, 01_ddim_inversion.ipynb.
when I use stable-diffusion-2-1(768ema), something wrong happened. It seems that 768ema can't correctly sample noise and inverse image. but 2-1-base works well. Is there any differences between them? How can I use stable-diffusion-2-1(768ema)? the cause I studying is unit4, 01_ddim_inversion.ipynb.
