Improving Diffusion-Based Image Synthesis with Context Prediction
Open paper
Benchmark
Model
Rank
Results
image-inpainting-on-celeba
ConPreDiff
–
LPIPS: 0.022
text-to-image-generation-on-coco
ConPreDiff
–
FID: 6.21
Zero shot FID: 6.21
Rank counts only results with a code link.