Fixed auto_segmentation app (cvat-ai#1562)

* disabled tf eager execution for auto_segmentation
* Update CHANGELOG.md

Author: azhavoro

CHANGELOG.md

@@ -46,6 +46,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Fixed an issue with large number of instances in instance mask (https://github.com/opencv/cvat/issues/1539)
 - Fixed full COCO dataset import error with conflicting labels in keypoints and detection (https://github.com/opencv/cvat/pull/1548)
 - Fixed COCO keypoints skeleton parsing and saving (https://github.com/opencv/cvat/issues/1539)
+- `tf.placeholder() is not compatible with eager execution` exception for auto_segmentation (https://github.com/opencv/cvat/pull/1562)
 
 ### Security
 -

cvat/apps/auto_segmentation/views.py

@@ -41,6 +41,11 @@ def _convert_to_segmentation(mask):
 
     ## INITIALIZATION
 
+    # workarround for tf.placeholder() is not compatible with eager execution
+    # https://github.com/tensorflow/tensorflow/issues/18165
+    import tensorflow as tf
+    tf.compat.v1.disable_eager_execution()
+
     # Root directory of the project
     ROOT_DIR = os.environ.get('AUTO_SEGMENTATION_PATH')
     # Import Mask RCNN