Fetching the paper…

InitNO: Boosting Text-to-Image Diffusion Models via Initial Noise Optimization · Around