>>107056742
There's no comfy workflow, and the model seems like some experimental half trained model, what is there to talk about?