TrojanZoo documentation

TrojanZoo is a universal pytorch platform for conducting security researches (especially for backdoor attacks/defenses) for image classification in deep learning.

Tutorials

• basic

TrojanZoo Documentation

• trojanzoo
  • attacks
  • configs
  • datasets
  • defenses
  • environ
  • models
  • optim
  • trainer
  • utils
    • data
    • fim
    • logger
    • model
    • module
    • tensor
    • train
• trojanvision
  • attacks
    • adv
    • backdoor
      • normal
      • clean_label
      • dynamic
  • configs
  • datasets
    • normal
    • folder
  • defenses
    • adv
    • backdoor
      • attack_agnostic
      • input_filtering
      • training_filtering
      • model_inspection
  • environ
  • marks
  • models
    • torchvision
    • nas
    • others
  • optim
  • trainer
  • utils

Indices and tables

• Index

• Module Index