There is a newer version of the record available.

Published May 15, 2023 | Version v0.2.0
Software Open

Adversarial Library

  • 1. ÉTS Montréal

Description

  • Add several segmentation attacks in the adv_lib.attacks.segmentation module: DAG, ASMA, PDGD, PDPGD and ALMA prox
  • Fix save method for adv_lib.utils.visdom_logger.VisdomLogger
  • Improve pytorch<1.9 compatibility
  • Allow tensor ε for PGD
  • Several minor improvements for FAB, DDN and FMN
  • Minor performance improvements using torch.amax instead of torch.max(dim=1).values
  • Fix FGA and VFGA best index selection
  • Improve README

Notes

If you use this software, please cite it using the metadata from this file.

Files

jeromerony/adversarial-library-v0.2.0.zip

Files (88.8 kB)

Name Size Download all
md5:0b3fa57e56b461813e602b74affb1a85
88.8 kB Preview Download

Additional details