Is there a 128 Clip training model?

I am not a developer so please forgive me if I am making a stupid question. My issue is that the mouth mask or size of it is too big so it is very noticeable. Is there a way I can run a smaller mouth size? I am suspecting 128 will work best. Or how can I make this work? 