KSNet-for-VSR

The repository contains the official implementation of "Kernel Dimension Matters: to Activate Avaliable Kernels for Real-time Video Super-resolution"

Introduction

We have updated the core implementation of KSNet. To start up, you can follow the instructions from BasicSR. More detailed introduction will be updated in a few weeks.

paper

Installation

Please follow installation guide.

Data Preparation

Please follow Dataset preparation.

Usage

Training Commands

Single GPU Training

PYTHONPATH="./:${PYTHONPATH}" \
CUDA_VISIBLE_DEVICES=0 \
python basicsr/train.py -opt options/train/ECBVSR/train_ECBVSR_REDS.yml

Distributed Training

PYTHONPATH="./:${PYTHONPATH}" \
CUDA_VISIBLE_DEVICES={GPU device id} \
python -m torch.distributed.launch --nproc_per_node={GPU device number} --master_port=4321 basicsr/train.py options/train/ECBVSR/train_ECBVSR_REDS.yml --launcher pytorch

or

CUDA_VISIBLE_DEVICES={GPU device id} \
./scripts/dist_train.sh {GPU device number} options/train/ECBVSR/train_ECBVSR_REDS.yml

TODOs

• Build baseline based on BasicSR
• Implement kernel split module
• Implement flow-guided DCN module
• Clean unrelated code
• Write clear instructions

Citation

If you find this work helpful, please consider citing:

@inproceedings{jin2023kernel,
  title={Kernel dimension matters: To activate available kernels for real-time video super-resolution},
  author={Jin, Shuo and Liu, Meiqin and Yao, Chao and Lin, Chunyu and Zhao, Yao},
  booktitle={Proceedings of the 31st ACM International Conference on Multimedia},
  pages={8617--8625},
  year={2023}
}