Fetching the paper…

LMFusion: Adapting Pretrained Language Models for Multimodal Generation · Around