🌋 LLaVA is a novel end-to-end trained large multimodal model that combines a vision encoder and Vicuna for general-purpose visual and language understanding. Updated to version 1.6.

7B 13B 34B

223.2K Pulls Updated 3 months ago

98 Tags