Lixin Gu

Latest

• ParVL: Parallel Scaling and Expandable Compute Allocation for Multimodal LLMs [preprint]