I realized when training the model with your script, my images output as 384 x 384 px . I have not yet used Diffusers directly for inference before, but from my understanding the model size for SD 2-0 depth is at least 512 px. How can I control this during training?
I realized when training the model with your script, my images output as 384 x 384 px . I have not yet used Diffusers directly for inference before, but from my understanding the model size for SD 2-0 depth is at least 512 px. How can I control this during training?