Support for quantizing this model: llama-joycaption-beta-one-hf-llava

[NeoChen1024]

Huggingface: https://huggingface.co/fancyfeast/llama-joycaption-beta-one-hf-llava GitHub: https://github.com/fpgaminer/joycaption

It's llava-based, but it doesn't specify "image_processor" like normal llava models.