Bug: `import tensorflow_hub` hits `AttributeError: module 'tensorflow._api.v2.compat.v2.__internal__' has no attribute 'register_load_context_function'`

TensorFlow
1

Python 3.13,

tensorflow==2.20.0
tensorflow-hub==0.16.1

import tensorflow_hub as hub

hits the following error:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[5], line 3
      1 import tensorflow as tf
      2 from tensorflow import keras
----> 3 import tensorflow_hub as hub

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tensorflow_hub/__init__.py:111
    103       print(
    104           "\n\nFailed to import tf_keras. Please note that tf_keras is not"
    105           " installed by default when you install tensorflow_hub. This is so"
    106           " that users can decide which tf_keras package to use. To use"
    107           " tensorflow_hub, please install a current version of tf_keras.\n\n"
    108       )
    109       raise
--> 111 _ensure_keras_2_importable()
    114 # pylint: disable=g-import-not-at-top
    115 # pylint: disable=g-bad-import-order
    116 # Symbols exposed via tensorflow_hub.
    117 from tensorflow_hub.keras_layer import KerasLayer

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tensorflow_hub/__init__.py:100, in _ensure_keras_2_importable()
     98 if version_fn and version_fn().startswith("3."):
     99   try:
--> 100     import tf_keras as keras
    101   except ImportError:
    102     # Print more informative error message, then reraise.
    103     print(
    104         "\n\nFailed to import tf_keras. Please note that tf_keras is not"
    105         " installed by default when you install tensorflow_hub. This is so"
    106         " that users can decide which tf_keras package to use. To use"
    107         " tensorflow_hub, please install a current version of tf_keras.\n\n"
    108     )

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/__init__.py:3
      1 """AUTOGENERATED. DO NOT EDIT."""
----> 3 from tf_keras import __internal__
      4 from tf_keras import activations
      5 from tf_keras import applications

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/__internal__/__init__.py:6
      4 from tf_keras.__internal__ import layers
      5 from tf_keras.__internal__ import losses
----> 6 from tf_keras.__internal__ import models
      7 from tf_keras.__internal__ import optimizers
      8 from tf_keras.__internal__ import utils

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/__internal__/models/__init__.py:3
      1 """AUTOGENERATED. DO NOT EDIT."""
----> 3 from tf_keras.src.models.cloning import clone_and_build_model
      4 from tf_keras.src.models.cloning import in_place_subclassed_model_state_restoration

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/__init__.py:21
      1 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
      2 #
      3 # Licensed under the Apache License, Version 2.0 (the "License");
   (...)     13 # limitations under the License.
     14 # ==============================================================================
     15 """Implementation of the TF-Keras API, the high-level API of TensorFlow.
     16 
     17 Detailed documentation and user guides are available at
     18 [keras.io](https://keras.io).
     19 """
---> 21 from tf_keras.src import applications
     22 from tf_keras.src import distribute
     23 from tf_keras.src import models

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/applications/__init__.py:18
      1 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
      2 #
      3 # Licensed under the Apache License, Version 2.0 (the "License");
   (...)     13 # limitations under the License.
     14 # ==============================================================================
     15 """Keras Applications are premade architectures with pre-trained weights."""
---> 18 from tf_keras.src.applications.convnext import ConvNeXtBase
     19 from tf_keras.src.applications.convnext import ConvNeXtLarge
     20 from tf_keras.src.applications.convnext import ConvNeXtSmall

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/applications/convnext.py:33
     31 from tf_keras.src import utils
     32 from tf_keras.src.applications import imagenet_utils
---> 33 from tf_keras.src.engine import sequential
     34 from tf_keras.src.engine import training as training_lib
     36 # isort: off

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/engine/sequential.py:24
     22 from tf_keras.src import layers as layer_module
     23 from tf_keras.src.engine import base_layer
---> 24 from tf_keras.src.engine import functional
     25 from tf_keras.src.engine import input_layer
     26 from tf_keras.src.engine import training

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/engine/functional.py:33
     31 from tf_keras.src.engine import input_spec
     32 from tf_keras.src.engine import node as node_module
---> 33 from tf_keras.src.engine import training as training_lib
     34 from tf_keras.src.engine import training_utils
     35 from tf_keras.src.saving import serialization_lib

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/engine/training.py:48
     46 from tf_keras.src.optimizers import optimizer_v1
     47 from tf_keras.src.saving import pickle_utils
---> 48 from tf_keras.src.saving import saving_api
     49 from tf_keras.src.saving import saving_lib
     50 from tf_keras.src.saving import serialization_lib

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/saving/saving_api.py:25
     22 from tensorflow.python.util.tf_export import keras_export
     24 from tf_keras.src.saving import saving_lib
---> 25 from tf_keras.src.saving.legacy import save as legacy_sm_saving_lib
     26 from tf_keras.src.utils import io_utils
     28 try:

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/saving/legacy/save.py:27
     25 from tf_keras.src.saving.legacy import serialization
     26 from tf_keras.src.saving.legacy.saved_model import load as saved_model_load
---> 27 from tf_keras.src.saving.legacy.saved_model import load_context
     28 from tf_keras.src.saving.legacy.saved_model import save as saved_model_save
     29 from tf_keras.src.saving.legacy.saved_model.utils import keras_option_scope

File ~/.local/share/virtualenvs/JupyterNotebooks-uVG1pv5y/lib/python3.13/site-packages/tf_keras/src/saving/legacy/saved_model/load_context.py:68
     64     """Returns whether under a load context."""
     65     return _load_context.in_load_context()
---> 68 tf.__internal__.register_load_context_function(in_load_context)

AttributeError: module 'tensorflow._api.v2.compat.v2.__internal__' has no attribute 'register_load_context_function'
3

Hi @khteh, Thanks for reporting the issue.
It looks like there is an incompatibility issue with your library versions. Please, Make sure your TensorFlow and tf-keras packages are the same version. Thank you!

4

I don’t cherry-pick the versions of any one of these libraries!

5

Could you please confirm whether you are importing tf-keras in your code without installing it first in your environment?

6

Does your question make sense? How possible is that?

7

I have tf-keras = "*" in Pipfile and don’t have any import of this module in my code at all

9

Using your specified versions, I was able to successfully import tensorflow_hub in Google Colab.

11

Are you sure you are using Python 3.13?

12

I’m running on Python 3.12.

13

And you said “Using your specified versions…”?

14

My apologies for the confusion. I am running on Python 3.12, not 3.13. Thank you!

15
17

Hi @khteh, Have you tried with Python 3.12?

19

No but Python 3.13.3

21

Hi @khteh, Could you please try downgrading Python 3.12, if possible? It appears to be incompatible with Python 3.13 and TensorFlow Hub.