Code for Enhancing Detail Preservation for Customized Text-to-Image Generation: A Regularization-Free Approach

ProFusion

Code for Enhancing Detail Preservation for Customized Text-to-Image Generation: A Regularization-Free Approach.

ProFusion is a framework for customizing pre-trained large-scale text-to-image generation models, which is Stable Diffusion 2 in our examples.

With ProFusion, you can generate infinite number of creative images for a novel/unique concept, with single testing image, on single GPU (~20GB are needed when fine-tune with batch size 1).

Example

• Install dependencies (we revised original diffusers);

    cd ./diffusers
    pip install -e .
    cd ..
    pip install accelerate==0.16.0 torchvision transformers>=4.25.1 datasets ftfy tensorboard Jinja2 regex tqdm joblib 
  

• Initialize Accelerate;

    accelerate config
  

• Download a model pre-trained on FFHQ;

• Customize model with a testing image, example is shown in the notebook test.ipynb;

Train Your Own Encoder

If you want to train a PromptNet encoder for other domains, or on your own dataset.

• First, prepare an image-only dataset;

  • In our example, we use FFHQ. Our pre-processed FFHQ can be found at google drive link.

• Then, run

    accelerate launch --mixed_precision="fp16" train.py\
          --pretrained_model_name_or_path="stabilityai/stable-diffusion-2-base" \
          --train_data_dir=./images_512 \
          --max_train_steps=80000 \
          --learning_rate=2e-05 \
          --output_dir="./promptnet" \
          --train_batch_size=8 \
          --promptnet_l2_reg=0.000 \
          --gradient_checkpointing
  

Citation

@misc{zhou2023enhancing,
  title={Enhancing Detail Preservation for Customized Text-to-Image Generation: A Regularization-Free Approach}, 
  author={Yufan Zhou and Ruiyi Zhang and Tong Sun and Jinhui Xu},
  year={2023},
  eprint={2305.13579},
  archivePrefix={arXiv},
  primaryClass={cs.CV}
}