we present Faster And More Efficient Subject Image Generation for Text-to-Image Diffusion Models, we introduce a new topic-customized method that requires no repeated training. It trains a plugand-play image prompt adapter with only 417M parameters, lightweight yet powerful, surpassing existing models in both text and image consistency. The DDCA inference code requires high memory consumption for visualizing the attention map and copy pipe, necessitating a V100. Removing these allows it to run inference on a 3090.

• [2024/5/1] 🔥 We release the code and models.

git clone  https://github.com/YZBPXX/DDCA
cd DDCA

pip install -r requirements.txt
# download the models
git lfs install
git clone https://huggingface.co/YZBPXX/CCDA

you can download models from here.

• demo: image generation with image prompt.

For training, you should install accelerate and make your own dataset into a pandas file.

bash train_sdxl.sh