search

wyhuai / DDNM

[ICLR 2023 Oral] Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model
MIT License
1.16k stars 83 forks source link

4x - SR for Imagenet doesn't give sharp results. Am i missing something? [ T_sampling: 300 travel_length: 1 travel_repeat: 1] #32

Open shreshthsaini opened 1 year ago

wyhuai commented 1 year ago

Sorry for the late reply.

The generative power of the unconditional ImageNet model is weak, and thus cannot provide sharp results.

You can try hq_demo for high-quality results, using a conditional ImageNet model.