python 3.x - module 'tensorflow_core.compat.v2' has no attribute '__internal__'

Question

I am using python 3.6.9 , keras 2.3.1, and (tf 2.0.0 or tf 2.1.0). Got the compatibility versions from this table.

When I import keras I get this error.

Answer 1

I was able to replicate your issue with the below combination as shown below

!pip install tensorflow==2.1.0
!pip install keras==2.3.1
import keras

Output:

Using TensorFlow backend.
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-3-88d96843a926> in <module>()
----> 1 import keras

8 frames
/usr/local/lib/python3.7/dist-packages/keras/initializers/__init__.py in populate_deserializable_objects()
     47 
     48   LOCAL.ALL_OBJECTS = {}
---> 49   LOCAL.GENERATED_WITH_V2 = tf.__internal__.tf2.enabled()
     50 
     51   # Compatibility aliases (need to exist in both V1 and V2).

AttributeError: module 'tensorflow_core.compat.v2' has no attribute '__internal__'

Fixed code:

Here issue is due to incompatibility between TF2.1 and Keras 2.3.1. Your issue can be resolved in two ways

1. Upgrade to Keras2.5.0rc0 and import keras
2. import keras from tensorflow as from tensorflow import keras