Faster way to compress vision-language training data into smaller datasets

Original: Multimodal Distribution Matching for Vision-Language Dataset Distillation

Source: arxiv.org ↗