Nvidia nvlm support

NVLM? Link?

oh -- we just merged a multimodal huggingface llm class, so it might be supported there. Not sure yet tbh though

Its not supported by ollama either, which is annoying

its also 72B, so you'll need a loooot of GPU

oh yes was hoping for a load_in_4bit option 🙂

ah yea, huggingface will probably let you do that (probably)