ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting (NeurIPS 2023, Spotlight)

Zongsheng Yue, Jianyi Wang, Chen Change Loy

Paper | Project Page | Video

⭐ If ResShift is helpful to your images or projects, please help star this repo. Thanks! 🤗

---

Diffusion-based image super-resolution (SR) methods are mainly limited by the low inference speed due to the requirements of hundreds or even thousands of sampling steps. Existing acceleration sampling techniques inevitably sacrifice performance to some extent, leading to over-blurry SR results. To address this issue, we propose a novel and efficient diffusion model for SR that significantly reduces the number of diffusion steps, thereby eliminating the need for post-acceleration during inference and its associated performance deterioration. Our method constructs a Markov chain that transfers between the high-resolution image and the low-resolution image by shifting the residual between them, substantially improving the transition efficiency. Additionally, an elaborate noise schedule is developed to flexibly control the shifting speed and the noise strength during the diffusion process. Extensive experiments demonstrate that the proposed method obtains superior or at least comparable performance to current state-of-the-art methods on both synthetic and real-world datasets, even only with 15 sampling steps.

---

Update

• 2023.08.15: Add .
• 2023.08.15: Add Gradio Demo.
• 2023.08.14: Add bicubic (matlab resize) model.
• 2023.08.14: Add Project Page.
• 2023.08.02: Add Replicate demo .
• 2023.07.31: Add Colab demo .
• 2023.07.24: Create this repo.

Requirements

• Python 3.9.16, Pytorch 1.12.1, xformers 0.0.20
• More detail (See environment.yaml) A suitable conda environment named ResShift can be created and activated with:

conda env create -f environment.yaml
conda activate ResShift

Applications

👉 Real-world image super-resolution

Online Demo

You can try our method through an online demo:

CUDA_VISIBLE_DEVICES=gpu_id python app.py

Inference

🐯 Real-world image super-resolution

CUDA_VISIBLE_DEVICES=gpu_id python inference_resshift.py -i [image folder/image path] -o [result folder] --task realsrx4 --chop_size 512

🦁 Bicubic (resize by Opencv) image super-resolution

CUDA_VISIBLE_DEVICES=gpu_id python inference_resshift.py -i [image folder/image path] -o [result folder] --task bicsrx4_cv2 --chop_size 512

🦁 Bicubic (resize by Matlab) image super-resolution

CUDA_VISIBLE_DEVICES=gpu_id python inference_resshift.py -i [image folder/image path] -o [result folder] --task bicsrx4_matlab --chop_size 512

Training

🐢 Prepare data

Download the training data and add the data path to the config file (data.train.params.dir_path or data.train.params.txt_file_path). To synthesize the testing dataset utilized in our paper, please refer to these scripts.

• Real-world and Bicubic image super-resolution: ImageNet and FFHQ (resized to 256x256)

🐬 Real-world Image Super-resolution

CUDA_VISIBLE_DEVICES=0,1,2,3 torchrun --standalone --nproc_per_node=4 --nnodes=1 main.py --cfg_path configs/realsr_swinunet_realesrgan256.yaml --save_dir [Logging Folder] --steps 15

🐳 Bicubic Image Super-resolution

CUDA_VISIBLE_DEVICES=0,1,2,3 torchrun --standalone --nproc_per_node=4 --nnodes=1 main.py --cfg_path configs/bicubic_swinunet_bicubic256.yaml --save_dir [Logging Folder]  --steps 15

License

This project is licensed under NTU S-Lab License 1.0. Redistribution and use should follow this license.

Acknowledgement

This project is based on Improved Diffusion Model, LDM, and BasicSR. We also adopt Real-ESRGAN to synthesize the training data for real-world super-resolution. Thanks for their awesome works.

Contact

If you have any questions, please feel free to contact me via [email protected].