Fetching the paper…

VLMo: Unified Vision-Language Pre-Training with Mixture-of-Modality-Experts · Around