Skip to content

shanshuo/ML-MWN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ML-MWN

Official code for the Multi-Label Meta Weighting for Long-Tailed Dynamic Scene Graph Generation (ICMR 2023) by Shuo Chen, Yingjun Du, Pascal Mettes, and Cees Snoek.

If you use this code for a paper, please cite:

@inproceedings{Chen2023Multi,
  author       = {Shuo Chen and
                  Ying{-}Jun Du and
                  Pascal Mettes and
                  Cees G. M. Snoek},
  title        = {Multi-Label Meta Weighting for Long-Tailed Dynamic Scene Graph Generation},
  booktitle    = {{ICMR}},
  year         = {2023},
}

Requirements

PyTorch 1.1.0+

Training

python meta_train.py -mode predcls -data_path <action_genome_path> -nepoch 20 -bce_loss -optimizer sgd -lr 0.001

Evaluation

python meta_test.py -mode predcls -data_path /<action_genome_path> -model_path <trained_model_path>

Acknowledgments

This repo is based on STTran, Meta-Weight-Net_Code-Optimization and meta-weight-net. We thank the authors for their work.

About

Official code for the Multi-Label Meta Weighting for Long-Tailed Dynamic Scene Graph Generation (ICMR 2023).

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published