I changed the base model to Schnell.
It works, trust me.
Just have to wait a few minutes until the model is loaded again, and the generations on the widget are in a few seconds.

iliketoasters changed pull request status to merged

Sign up or log in to comment