Accelerated Multi-Modal MR Imaging with Transformers

Dependencies

• numpy==1.18.5
• scikit_image==0.16.2
• torchvision==0.8.1
• torch==1.7.0
• runstats==1.8.0
• pytorch_lightning==1.0.6
• h5py==2.10.0
• PyYAML==5.4

:fire: NEWS :fire:

• We have uploaded the csv.files of the paired data.

multi gpu train

python -m torch.distributed.launch --nproc_per_node=8   train.py --experiment sr_multi_cross

single gpu train

python train.py --experiment sr_multi_cross

multi gpu test

python -m torch.distributed.launch --nproc_per_node=8   test.py --experiment sr_multi_cross

single gpu test

python test.py --experiment sr_multi_cross

--experiment is the experiment you running. And you can change the config file to set the parameter when training or testing.

Citation

@article{feng2021accelerated,
  title={Accelerated Multi-Modal MR Imaging with Transformers},
  author={Feng, Chun-Mei and Yan, Yunlu and Chen, Geng and Fu, Huazhu and Xu, Yong and Shao, Ling},
  journal={arXiv e-prints},
  pages={arXiv--2106},
  year={2021}
}