Official Implementation of LaViDa: :A Large Diffusion Language Model for Multimodal Understanding - View it on GitHub