How to use the Deep Hat v1 open source model running on Hugging Face’s inference provider service using Silly Tavern