First concept LORA, meant to generate images of the viewer about to be gagged with a piece of cloth. From my testing it should work fine with other LORAs and with different styles.
Update 1.5: Retrained with larger dataset, from my testing it should give better results now