Preprint

OLA-VLM: Elevating Perception in Multimodal LLMs with Auxiliary Embedding Distillation