multimodal

Work involving models that combine text with images, audio, video, or other modalities in training, inference, or evaluation.