StyleEntity

This repository contains the code for the paper "Named Entity Driven Zero-Shot Image Manipulation" presented at CVPR 2024.

Prerequisites

To run the code in this repository, ensure you have the following environment setup:

• torch>=1.6.0
• BasicSR

Preparing Pre-trained Models

1. Download the StyleGAN2 checkpoint stylegan2_ffhq_config_f_1024_official-3ab41b38.pth from the BasicSR repository: BasicSR Checkpoints.
2. Save the downloaded checkpoint in an accessible directory for later use.

Encoding Named Entities

Before training, you need to encode the Named Entity texts. Run the following script to perform the encoding:

python scripts/encode_nes.py

Training

To train the model, use the following command with the specified training configuration file:

python train.py -opt options/train.yaml

Pre-trained Models

You can download pre-trained models from the following link:

• StyleEntity Pre-trained Models

Citation

If you find this repository useful in your research, please consider citing our paper:

@inproceedings{StyleEntity2024,
    author    = {Feng, Zhida and Chen, Li and Tian, Jing and Liu, JiaXiang and Feng, Shikun},
    title     = {Named Entity Driven Zero-Shot Image Manipulation},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2024},
    pages     = {9110-9119}
}

License

This repository is released under the MIT License. See the LICENSE file for more details.

Contact

For any questions or issues, please open an issue on this repository or contact feng.zhida@outlook.com.