New Features and Improvements

• Expectation Over Transformation Wrapper
• Add MIFGSM Attack

New Features and Improvements

• Fixed a bug that AdamPGD attacks actually didn't use the Adam optimizer
• Attacks now verify that the input to them lies within the bounds of the model

New Features and Improvements

• Added AdamPGD attack
• Added pointwise attack
• Hop-skip-jump attack bug fix (thanks @zhuangzi926)
• other improvements and bug fixes

Serious bug fix for SaltAndPepperAttack (thanks @maurapintor @zangobot)

New Features and Improvements

• PGD now supports targeted attacks (thanks @zimmerrol)
• DDN attack bug fixes (thanks @maurapintor)
• Brendel Bethge attack bug fixes (thanks @wielandbrendel)
• other improvements and bug fixes

Foolbox release for Zenodo

• added our JOSS paper
• added a performance comparison between Foolbox 1, 2, and 3
• improved tests
• fixed the TensorFlow example code
• improved examples
• improved tutorial
• updated dependencies

Bug fix

New Features

• ported HopSkipJump attack to v3
• added clipping-aware noise attacks
• model wrappers now support data_format
• JAXModel now supports data_format
• improved documentation

Bug Fixes

• EADAttack bug fixes
• GenAttack bug fixes
• Other bug fixes and improvements

Fixed the version number

Fixes a bug in the BrendelBethgeAttack and updated Numba to silence warnings.

Fixes a bug in the BrendelBethgeAttack (thanks @AidanKelley)

Bug fixes

• type annotations are now correctly exposed using py.typed (file was missing in MANIFEST)
• TransformBoundsWrapper now correctly handles data_format (thanks @zimmerrol)

New Features

Foolbox 3 aka Foolbox Native has been rewritten from scratch with performance in mind. All code is running natively in PyTorch, TensorFlow and JAX, and all attacks have been rewritten with real batch support.

New Features

• added foolbox.gradient_estimators
• improved attack hyperparameter documentation

Foolbox 3 aka Foolbox Native has been rewritten from scratch with performance in mind. All code is running natively in PyTorch, TensorFlow and JAX, and all attacks have been rewritten with real batch support.

Warning: This is a pre-release beta version. Expect breaking changes.

New Features

• fixed PyTorch model gradients (fixes DeepFool with batch size > 1)
• added support for TensorFlow 2.0 and newer (Graph and Eager mode)
• refactored the tests
• support for the latest randomgen version

New Features

• new EnsembleAveragedModel (thanks to @zimmerrol)
• new foolbox.utils.flatten
• new foolbox.utils.atleast_kd
• new foolbox.utils.accuracy
• PyTorchModel now always warns if model is in train mode, not just once
• batch support for ModelWithEstimatedGradients

Bug fixes

• fixed dtype when using Adam PGD with a PyTorch model
• fixed CW attack hyperparameters

New Features

• support for Foolbox extensions using the foolbox.ext namespace

New Features

• New foolbox.models.JAXModel class to support JAX models (https://github.com/google/jax)
• The preprocessing argument of models now supports a flip_axis key to support common preprocessing operations like RGB to BGR in a nice way. This builds on the ability to pass dicts to preprocessing introduced in Foolbox 2.0.

Bug fixes and improvements

• Fixed a serious bug in the LocalSearchAttack (thanks to @duoergun0729)
• foolbox.utils.samples now warns if samples are repeated
• foolbox.utils.sampels now uses PNGs instead of JPGs (except for ImageNet)
• Other bug fixes
• Improved docstrings
• Improved docs