The dataset focuses on hierarchical hate detection in Bangla memes with multi-level annotations across binary labels, hate categories, and targeted groups.
It contains 3,819 annotated meme samples collected from Facebook (2,517) and Instagram (1,302) between April 2022 ā May 2025.
Binary classification: Hate (1,769) vs.⦠See the full description on the dataset page:
https://huggingface.co/datasets/Ayon128/BanHateMe.