You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fixed bug in PyTorchClassifier.get_activations to also apply preprocessing if argument framework=True. This fix likely changes the results obtained with BullseyePolytopeAttackPyTorch, the main attack using framework=True. (Replace x with x_preprocessed in PyTorchClassifier.get_activations #1471)
This discussion was created from the release ART 1.9.1.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
This release of ART 1.9.1 provides updates to ART 1.9.
Added
KerasClassifier.compute_loss
. (Add support in KerasClassifier.compute_loss for TF 1.15 as backend #1466)art.defences.preprocessor.VideoCompression*
. (Preprocessing defences in PyTorchGoturn and VideoCompression for [0, 1] range #1470)Changed
[None]
Removed
[None]
Fixed
art.utils.load_nursery
for loading nursery dataset with argumentraw=True
. (Fix bug when loading nursery data #1460)matplotlib
to keep it an optional dependency. (Fix import of matplotlib in laser attack utils #1467)PyTorchGoturn.predict
by adding back missing sample dimension. (Preprocessing defences in PyTorchGoturn and VideoCompression for [0, 1] range #1470)PyTorchClassifier.get_activations
to also apply preprocessing if argumentframework=True
. This fix likely changes the results obtained withBullseyePolytopeAttackPyTorch
, the main attack usingframework=True
. (Replace x with x_preprocessed in PyTorchClassifier.get_activations #1471)This discussion was created from the release ART 1.9.1.
Beta Was this translation helpful? Give feedback.
All reactions