Hybrid Diffusion-Transformer Models

U-ViT

Variant of Vision Transformer where U-Net connections are integrated to effectively combine multi-scale features in diffusion models.

← Tillbaka