search

cyclomon / DiffuseIT

Official repository of "Diffusion-based Image Translation using Disentangled Style and Content Representation" ( ICLR 2023 )
MIT License
257 stars 15 forks source link

what do you mean about face identity loss in github? #12

Closed justinday123 closed 11 months ago

justinday123 commented 11 months ago

hi, thank you for your work. I want to know about identity loss you wrote in github. "For face identity loss when using FFHQ pre-trained model, download pre-trained ArcFace model LINK"

justinday123 commented 11 months ago

Also, did you used DINO-vit which is pretrained with imagenet dataset?

cyclomon commented 11 months ago

Hi, face identity loss is to maintain the face identity between source face image and translated face image. We use pre-trained face identity embedding model to implement.

Also, we used DINO-vit pre-trained with imagenet.