Jan, 2024
改善基于扩散的图像合成
Improving Diffusion-Based Image Synthesis with Context Prediction
Ling Yang, Jingwei Liu, Shenda Hong, Zhilong Zhang, Zhilin Huang...
TL;DRDiffusion models for image synthesis are improved using ConPreDiff, a method that incorporates context prediction, achieving superior results in unconditional image generation, text-to-image generation, and image inpainting.