metadata
pipeline_tag: image-text-to-text
This model is the stage 2 checkpoint of one of the thirteen settings, CLIP+DINOv2@336, used in the Law of Vision Representation in MLLMs.
pipeline_tag: image-text-to-text
This model is the stage 2 checkpoint of one of the thirteen settings, CLIP+DINOv2@336, used in the Law of Vision Representation in MLLMs.