Multimodal memes across two binary hate speech classes. Content focuses on multimodal reasoning where both image and text must be processed together to identify hateful intent.
Use Cases
- Train a multimodal classifier using image and text features to predict hate speech labels
- Evaluate model performance on multimodal reasoning tasks
- Research the integration of visual and textual signals for social media safety
Strengths
- Multimodal data format combining images and text
- Binary labels for hate speech detection
- Benchmark created by Facebook AI for multimodal reasoning