We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Is there a way for it to use the latent consistency model? Is it possible to run it on the CPU or use Kandinsky 3.0 as it is very fast to use