EVCap: Retrieval-Augmented Image Captioning with External Visual-Name Memory for Open-World Comprehension | Read Paper on Bytez