Official Implementation for Diffusion-Based Scene Graph to Image Generation with Masked Contrastive Pre-Training.

🚩 New Updates : We release LAION-SG, a large-scale dataset with high-quality structural annotations of scene graphs (SG), which precisely describe attributes and relationships of multiple objects, effectively representing the semantic structure in complex scenes. Based on LAION-SG, we also provide a new foundation model SDXL-SG to incorporate structural annotation information into the generation process.

git clone https://github.com/YangLing0818/SGDiff.git
cd SGDiff

conda env create -f sgdiff.yaml
conda activate sgdiff
mkdir pretrained

The instructions of data pre-processing can be found here.

Our masked contrastive pre-trained models of SG-image pairs for COCO and VG datasets are provided in here, please download them and put them in the 'pretrained' directory.

And the pretrained VQVAE for embedding image to latent can be obtained from https://ommer-lab.com/files/latent-diffusion/vq-f8.zip

The instructions of SG-image pretraining can be found in the folder "sg_image_pretraining/"

Kindly note that one should not skip the training stage and test directly. For single gpu, one can use

python trainer.py --base CONFIG_PATH -t --gpus 0,

NOT OFFICIAL: Alternatively, if you don't want to train the model from scratch you can download trained weights from the following link: VG weight, COCO weight

Checkpoint trained for only 150 epochs.

python testset_ddim_sampler.py

If you found the codes are useful, please cite our paper

@article{yang2022diffusionsg,
  title={Diffusion-based scene graph to image generation with masked contrastive pre-training},
  author={Yang, Ling and Huang, Zhilin and Song, Yang and Hong, Shenda and Li, Guohao and Zhang, Wentao and Cui, Bin and Ghanem, Bernard and Yang, Ming-Hsuan},
  journal={arXiv preprint arXiv:2211.11138},
  year={2022}
}

@article{li2024laion,
  title={LAION-SG: An Enhanced Large-Scale Dataset for Training Complex Image-Text Models with Structural Annotations},
  author={Li, Zejian and Meng, Chenye and Li, Yize and Yang, Ling and Zhang, Shengyuan and Ma, Jiarui and Li, Jiayi and Yang, Guang and Yang, Changyuan and Yang, Zhiyuan and others},
  journal={arXiv preprint arXiv:2412.08580},
  year={2024}
}