Remove TensorFlow2v1

Signed-off-by: Beat Buesser <[email protected]>

Author: beat-buesser

.github/workflows/ci-tensorflow-v2.yml

@@ -28,13 +28,6 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - name: TensorFlow 2.18.1v1 (Keras 3.10.0 Python 3.10)
-            framework: tensorflow2v1
-            python: '3.10'
-            tensorflow: 2.18.1
-            tf_version: v2
-            keras: 3.10.0
-            tf_addons: 0.23.0
           - name: TensorFlow 2.18.1 (Keras 3.10.0 Python 3.10)
             framework: tensorflow
             python: '3.10'

art/attacks/evasion/auto_conjugate_gradient.py

@@ -115,10 +115,7 @@ def __init__(
             "cross_entropy", or "difference_logits_ratio"
         :param verbose: Show progress bars.
         """
-        from art.estimators.classification import TensorFlowClassifier, TensorFlowV2Classifier, PyTorchClassifier
-
-        if isinstance(estimator, TensorFlowClassifier):
-            raise ValueError("This attack does not support TensorFlow  v1.")
+        from art.estimators.classification import TensorFlowV2Classifier, PyTorchClassifier
 
         if loss_type not in self._predefined_losses:
             raise ValueError(

art/attacks/evasion/rescaling_auto_conjugate_gradient.py

@@ -113,10 +113,7 @@ def __init__(
             "cross_entropy", or "difference_logits_ratio"
         :param verbose: Show progress bars.
         """
-        from art.estimators.classification import TensorFlowClassifier, TensorFlowV2Classifier, PyTorchClassifier
-
-        if isinstance(estimator, TensorFlowClassifier):
-            raise ValueError("This attack does not support TensorFlow  v1.")
+        from art.estimators.classification import TensorFlowV2Classifier, PyTorchClassifier
 
         if loss_type not in self._predefined_losses:
             raise ValueError(

art/data_generators.py

@@ -257,87 +257,6 @@ def get_batch(self) -> tuple:
         return tuple(batch)
 
 
-class TensorFlowDataGenerator(DataGenerator):  # pragma: no cover
-    """
-    Wrapper class on top of the TensorFlow native iterators :class:`tf.data.Iterator`.
-    """
-
-    def __init__(
-        self,
-        sess: "tf.Session",
-        iterator: "tf.data.Iterator",
-        iterator_type: str,
-        iterator_arg: dict | tuple | "tf.Operation",
-        size: int,
-        batch_size: int,
-    ) -> None:
-        """
-        Create a data generator wrapper for TensorFlow. Supported iterators: initializable, reinitializable, feedable.
-
-        :param sess: TensorFlow session.
-        :param iterator: Data iterator from TensorFlow.
-        :param iterator_type: Type of the iterator. Supported types: `initializable`, `reinitializable`, `feedable`.
-        :param iterator_arg: Argument to initialize the iterator. It is either a feed_dict used for the initializable
-        and feedable mode, or an init_op used for the reinitializable mode.
-        :param size: Total size of the dataset.
-        :param batch_size: Size of the minibatches.
-        :raises `TypeError`, `ValueError`: If input parameters are not valid.
-        """
-
-        import tensorflow.compat.v1 as tf
-
-        super().__init__(size=size, batch_size=batch_size)
-        self.sess = sess
-        self._iterator = iterator
-        self.iterator_type = iterator_type
-        self.iterator_arg = iterator_arg
-
-        if not isinstance(iterator, tf.data.Iterator):
-            raise TypeError("Only support object tf.data.Iterator")
-
-        if iterator_type == "initializable":
-            if not isinstance(iterator_arg, dict):
-                raise TypeError(f"Need to pass a dictionary for iterator type {iterator_type}")
-        elif iterator_type == "reinitializable":
-            if not isinstance(iterator_arg, tf.Operation):
-                raise TypeError(f"Need to pass a TensorFlow operation for iterator type {iterator_type}")
-        elif iterator_type == "feedable":
-            if not isinstance(iterator_arg, tuple):
-                raise TypeError(f"Need to pass a tuple for iterator type {iterator_type}")
-        else:
-            raise TypeError(f"Iterator type {iterator_type} not supported")
-
-    def get_batch(self) -> tuple:
-        """
-        Provide the next batch for training in the form of a tuple `(x, y)`. The generator should loop over the data
-        indefinitely.
-
-        :return: A tuple containing a batch of data `(x, y)`.
-        :raises `ValueError`: If the iterator has reached the end.
-        """
-        import tensorflow as tf
-
-        # Get next batch
-        next_batch = self.iterator.get_next()
-
-        # Process to get the batch
-        try:
-            if self.iterator_type in ("initializable", "reinitializable"):
-                return self.sess.run(next_batch)
-            return self.sess.run(next_batch, feed_dict=self.iterator_arg[1])
-        except (tf.errors.FailedPreconditionError, tf.errors.OutOfRangeError):
-            if self.iterator_type == "initializable":
-                self.sess.run(self.iterator.initializer, feed_dict=self.iterator_arg)
-                return self.sess.run(next_batch)
-
-            if self.iterator_type == "reinitializable":
-                self.sess.run(self.iterator_arg)
-                return self.sess.run(next_batch)
-
-            self.sess.run(self.iterator_arg[0].initializer)
-            return self.sess.run(next_batch, feed_dict=self.iterator_arg[1])
-
-
 class TensorFlowV2DataGenerator(DataGenerator):
     """
     Wrapper class on top of the TensorFlow v2 native iterators :class:`tf.data.Iterator`.

art/defences/preprocessor/__init__.py

@@ -10,7 +10,6 @@
 from art.defences.preprocessor.cutout.cutout_tensorflow import CutoutTensorFlowV2
 from art.defences.preprocessor.feature_squeezing import FeatureSqueezing
 from art.defences.preprocessor.gaussian_augmentation import GaussianAugmentation
-from art.defences.preprocessor.inverse_gan import DefenseGAN, InverseGAN
 from art.defences.preprocessor.jpeg_compression import JpegCompression
 from art.defences.preprocessor.label_smoothing import LabelSmoothing
 from art.defences.preprocessor.mixup.mixup import Mixup

art/defences/preprocessor/inverse_gan.py

@@ -12,11 +12,10 @@
 from art.estimators.keras import KerasEstimator
 from art.estimators.pytorch import PyTorchEstimator
 from art.estimators.scikitlearn import ScikitlearnEstimator
-from art.estimators.tensorflow import TensorFlowEstimator, TensorFlowV2Estimator
+from art.estimators.tensorflow import TensorFlowV2Estimator
 
 from art.estimators import certification
 from art.estimators import classification
-from art.estimators import encoding
 from art.estimators import generation
 from art.estimators import object_detection
 from art.estimators import poison_mitigation

art/estimators/__init__.py

@@ -20,9 +20,5 @@
 from art.estimators.classification.hugging_face import HuggingFaceClassifierPyTorch
 from art.estimators.classification.query_efficient_bb import QueryEfficientGradientEstimationClassifier
 from art.estimators.classification.scikitlearn import SklearnClassifier
-from art.estimators.classification.tensorflow import (
-    TFClassifier,
-    TensorFlowClassifier,
-    TensorFlowV2Classifier,
-)
+from art.estimators.classification.tensorflow import TensorFlowV2Classifier
 from art.estimators.classification.xgboost import XGBoostClassifier