Hi,
what settings should be used to reproduce the figures from paper?
I used:

python main.py --json_path resource/example.json --model_path "stabilityai/stable-diffusion-xl-base-1.0" 

but visually the results aren't as good: