LLaVa, a visual instruction tuned version of LLaMa and other large language models, can now be used natively with the Transformers library.
