AttriDiffuser: Text-to-Facial Attributes Synthesizing with Adversarial Enhanced Diffusion Model

Our pipeline

Environment

Follow ldm's environment configuration.

Or use our configuration file.

git clone https://github.com/sunmeng7/AttriDiffuser.git
cd AttriDiffuser
conda env create -f envs.yaml

Models

Baidu Yun

link : https://pan.baidu.com/s/12kekZQDsQT9-Nh_kD41lKg

code : 1qrw

Visual Results

Multimodal Voxceleb

CelebA-HQ

Acknowledgement

We have used codes from other great research work, including LDM, SDM, GLIGEN.

We sincerely thank these excellent work.

For details on the Face Diversity Discriminator, see our previous work TYS.