Article

Representation Alignment for Just Image Transformers is not Easier than You Think

Improving Video Diffusion Transformer Training by Multi-Feature Fusion and Alignment from Self-Supervised Vision Encoders