2024
SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis
ICLR 2024spotlight
We present Stable Diffusion XL (SDXL), a latent diffusion model for text-to-image synthesis. Compared to previous versions of Stable Diffusion, SDXL leverages a three times larger UNet backbone, achieved by significantly increasing the number of attention blocks and including a second text encoder.…