Datasets accompanying the paper "M5 – A Diverse Benchmark to Assess the Performance of Large Multimodal Models Across Multilingual and Multicultural Vision-Language Tasks" presented at EMNLP 2024 - View it on GitHub