suggested fixes

research/object_detection/meta_architectures/context_rcnn_lib_v2.py → research/object_detection/meta_architectures/context_rcnn_lib_tf2.py

@@ -18,7 +18,7 @@ from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 
-import tensorflow.compat.v1 as tf
+import tensorflow as tf
 
 # The negative value used in padding the invalid weights.
 _NEGATIVE_PADDING_VALUE = -100000
@@ -88,7 +88,7 @@ class AttentionBlock(tf.keras.layers.Layer):
     # box_features becomes [batch_size, max_num_proposals, channels].
     input_features = tf.reduce_mean(input_features, [2, 3])
 
-    with tf.variable_scope("AttentionBlock"):
+    with tf.name_scope("AttentionBlock"):
       queries = project_features(
           input_features, self._bottleneck_dimension, self._is_training,
           self._query_proj, normalize=True)

research/object_detection/meta_architectures/context_rcnn_lib_tf2_test.py

@@ -22,7 +22,7 @@ import unittest
 from absl.testing import parameterized
 import tensorflow.compat.v1 as tf
 
-from object_detection.meta_architectures import context_rcnn_lib_v2 as context_rcnn_lib
+from object_detection.meta_architectures import context_rcnn_lib_tf2 as context_rcnn_lib
 from object_detection.utils import test_case
 from object_detection.utils import tf_version
 

research/object_detection/meta_architectures/context_rcnn_meta_arch.py

@@ -27,7 +27,7 @@ import functools
 
 from object_detection.core import standard_fields as fields
 from object_detection.meta_architectures import context_rcnn_lib
-from object_detection.meta_architectures import context_rcnn_lib_v2
+from object_detection.meta_architectures import context_rcnn_lib_tf2
 from object_detection.meta_architectures import faster_rcnn_meta_arch
 from object_detection.utils import tf_version
 
@@ -272,7 +272,7 @@ class ContextRCNNMetaArch(faster_rcnn_meta_arch.FasterRCNNMetaArch):
           attention_temperature=attention_temperature,
           is_training=is_training)
     else:
-      self._context_feature_extract_fn = context_rcnn_lib_v2.AttentionBlock(
+      self._context_feature_extract_fn = context_rcnn_lib_tf2.AttentionBlock(
           bottleneck_dimension=attention_bottleneck_dimension,
           attention_temperature=attention_temperature,
           freeze_batchnorm=freeze_batchnorm,