Fetching the paper…

Re-ViLM: Retrieval-Augmented Visual Language Model for Zero and Few-Shot Image Captioning · Around