MoPE: Parameter-Efficient and Scalable Multimodal Fusion via Mixture of Prompt Experts

Open in new window