ArMeme is the first multimodal Arabic memes dataset that includes both text and images, collected from various social media platforms. It serves as the first resource dedicated to Arabic multimodal research. While the dataset has been annotated to identify propaganda in memes, it is versatile and can be utilized for a wide range of other research purposes, including sentiment analysis, hate speech detection, cultural studies, meme generation, and… See the full description on the dataset page:
https://huggingface.co/datasets/QCRI/ArMeme.
