aiml package

Subpackages

• aiml.attack package
  • Submodules
  • aiml.attack.adversarial_attacks module
    • auto_projected_cross_entropy()
    • auto_projected_difference_logits_ratio()
    • carlini_L0_attack()
    • carlini_L2_attack()
    • carlini_Linf_attack()
    • deep_fool_attack()
    • pixel_attack()
    • square_attack()
    • zoo_attack()
  • aiml.attack.attack_evaluation module
    • attack_evaluation()
  • Module contents
• aiml.evaluation package
  • Submodules
  • aiml.evaluation.check_accuracy module
    • check_accuracy()
    • check_accuracy_with_flags()
  • aiml.evaluation.dynamic module
    • decide_attack()
  • aiml.evaluation.evaluate module
    • evaluate()
  • Module contents
• aiml.load_data package
  • Submodules
  • aiml.load_data.generate_parameter module
    • generate_parameter()
  • aiml.load_data.load_model module
    • load_model()
  • aiml.load_data.load_set module
    • load_test_set()
    • load_train_set()
  • aiml.load_data.normalize_datasets module
    • check_normalize()
    • denormalize()
    • get_mean_std()
    • get_transforms()
    • normalize_and_check_datasets()
    • normalize_datasets()
  • Module contents
• aiml.surrogate_model package
  • Submodules
  • aiml.surrogate_model.create_surrogate_model module
    • create_substitute()
    • create_surrogate_model()
    • get_num_classes()
  • aiml.surrogate_model.models module
    • LogSoftmaxModule
      • LogSoftmaxModule.model
      • LogSoftmaxModule.forward()
      • LogSoftmaxModule.predict_step()
    • Surrogate
      • Surrogate.oracle
      • Surrogate.substitute
      • Surrogate.loss_fn
      • Surrogate.accuracy
      • Surrogate.configure_optimizers()
      • Surrogate.forward()
      • Surrogate.predict_step()
      • Surrogate.training_step()
      • Surrogate.validation_step()
    • create_substitute_model()
  • aiml.surrogate_model.utils module
    • choose_dataset()
    • find_clip_range()
    • get_data()
    • get_labels()
    • get_transforms()
    • inverse_normalize()
    • load_cifar10()
  • Module contents

Module contents

Adversarial Insight ML is a package that provides evaluation of ML models based on adversarial attacks.