README.md

SimTriplet

SimTriplet: PyTorch Implementation Paper

Data description

Image data used for model pretrain, finetune and test can be downloaded via Google drive link:

Train: link
Finetune: link
Test: link

The complete training and testing datasets will be provided once the paper is officially published.

Pre-train SimTriplet

main_tcga_mixpresition.py

--data_dir
../Data/
--log_dir
../logs/
-c
configs/TCGA_triple.yaml
--ckpt_dir
./checkpoint/

Finetune linear classifier

TCGA_linear_cross_val.py

Test

test_patho_prob_major_vote.py

Well trained model

Model pretrained on TCGA data is provided here

Citation

If you find this repository useful in your research, please cite:

@misc{liu2021simtriplet,
      title={SimTriplet: Simple Triplet Representation Learning with a Single GPU}, 
      author={Quan Liu and Peter C. Louis and Yuzhe Lu and Aadarsh Jha and Mengyang Zhao and Ruining Deng and Tianyuan Yao and Joseph T. Roland and Haichun Yang and Shilin Zhao and Lee E. Wheless and Yuankai Huo},
      year={2021},
      eprint={2103.05585},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}