RIVA: A Pre-trained Tweet Multimodal Model Based on Text-image Relation for Multimodal NER

4187

Knowledge Triples