AttributeError: module 'tensorflow' has no attribute 'compat' when loading tf.compat.v1.train.SessionRunHook

Question

I can see that this question has been asked before here tensorflow-has-no-attribute-compat

but the answer given was to

Microsoft Visual C++ 2015-2019 Redistributable (x64)

It did not work for the previous member it has not worked for me either. I have visual studio 2019 installed. I downloaded it anyways and ran a repair of (MV C++) just in case. Still getting same error.

So that being said I have not found a valid solution for this anywhere on google or stackoverflow. Here are some spec details of what I have installed.

tensorflow-gpu 2.1
python 3.7.7
CUDA 10.1
Anaconda 3.7

Looks like gpu started successfully.

2020-06-28 07:19:47.851257: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll

python packages for 3.7.7. All packages installed via pip. Only tensorflow installed via conda. Dont think it makes a difference. I used conda so I can specify install version 2.1, pip would automatically install 2.2.

Package                Version
---------------------- ----------------------------
absl-py                0.9.0
appdirs                1.4.4
astor                  0.7.1
astroid                2.4.2
astropy                4.0.1.post1
attrs                  19.3.0
backcall               0.2.0
bayesian-optimization  1.2.0
black                  19.10b0
bleach                 3.1.5
blinker                1.4
brotlipy               0.7.0
cachetools             4.1.0
certifi                2020.6.20
cffi                   1.14.0
chardet                3.0.4
click                  7.1.2
cloudpickle            1.3.0
colorama               0.4.3
confuse                1.3.0
cryptography           2.9.2
cycler                 0.10.0
decorator              4.4.2
defusedxml             0.6.0
entrypoints            0.3
future                 0.18.2
gast                   0.2.2
gitdb                  4.0.5
GitPython              3.1.3
google-auth            1.17.2
google-auth-oauthlib   0.4.1
google-pasta           0.2.0
grpcio                 1.27.2
gym                    0.17.2
h2o                    3.30.0.5
h5py                   2.10.0
htmlmin                0.1.12
idna                   2.10
ImageHash              4.1.0
importlib-metadata     1.7.0
invoke                 1.4.1
ipykernel              5.3.0
ipython                7.16.1
ipython-genutils       0.2.0
ipywidgets             7.5.1
isort                  4.3.21
jedi                   0.17.1
Jinja2                 2.11.2
joblib                 0.15.1
json5                  0.9.5
jsonschema             3.2.0
jupyter                1.0.0
jupyter-client         6.1.3
jupyter-console        6.1.0
jupyter-core           4.6.3
jupyterlab             2.1.5
jupyterlab-server      1.1.5
kaggle                 1.5.6
Keras                  2.4.3
Keras-Applications     1.0.8
Keras-Preprocessing    1.1.0
kiwisolver             1.2.0
lazy-object-proxy      1.4.3
lightgbm               2.3.1
llvmlite               0.33.0
Markdown               3.2.2
MarkupSafe             1.1.1
matplotlib             3.2.2
mccabe                 0.6.1
missingno              0.4.2
mistune                0.8.4
mkl-service            2.3.0
nbconvert              5.6.1
nbdime                 2.0.0
nbformat               5.0.7
networkx               2.4
notebook               6.0.3
numba                  0.50.1
numpy                  1.19.0
oauthlib               3.0.1
opt-einsum             0+untagged.56.g2664021.dirty
packaging              20.4
pandas                 1.0.5
pandas-profiling       2.8.0
pandocfilters          1.4.2
parso                  0.7.0
path                   13.1.0
path.py                12.4.0
pathspec               0.8.0
patsy                  0.5.1
phik                   0.10.0
pickleshare            0.7.5
Pillow                 7.1.2
pip                    20.1.1
plotly                 4.8.2
prometheus-client      0.8.0
prompt-toolkit         3.0.5
protobuf               3.12.3
py4j                   0.10.9
pyasn1                 0.4.8
pyasn1-modules         0.2.7
pycparser              2.20
pyglet                 1.5.0
Pygments               2.6.1
PyJWT                  1.7.1
pylint                 2.5.3
pyOpenSSL              19.1.0
pyparsing              2.4.7
pyreadline             2.1
pyrsistent             0.16.0
PySocks                1.7.1
pyspark                3.0.0
python-dateutil        2.8.1
python-slugify         4.0.0
pytz                   2020.1
PyWavelets             1.1.1
pywin32                228
pywinpty               0.5.7
PyYAML                 5.3.1
pyzmq                  19.0.1
qtconsole              4.7.5
QtPy                   1.9.0
regex                  2020.6.8
requests               2.24.0
requests-oauthlib      1.2.0
retrying               1.3.3
rsa                    4.6
scikit-learn           0.23.1
scipy                  1.5.0
seaborn                0.10.1
Send2Trash             1.5.0
setuptools             47.3.1.post20200616
six                    1.15.0
smmap                  3.0.4
statsmodels            0.11.1
tabulate               0.8.7
tangled-up-in-unicode  0.0.6
tensorboard            2.2.2
tensorboard-plugin-wit 1.6.0.post3
tensorflow             2.1.0
tensorflow-estimator   2.2.0
termcolor              1.1.0
terminado              0.8.3
testpath               0.4.4
text-unidecode         1.3
threadpoolctl          2.1.0
toml                   0.10.1
tornado                6.0.4
tqdm                   4.46.1
traitlets              4.3.3
typed-ast              1.4.1
urllib3                1.24.3
visions                0.4.4
wcwidth                0.2.5
webencodings           0.5.1
Werkzeug               0.16.1
wheel                  0.34.2
widgetsnbextension     3.5.1
win-inet-pton          1.1.0
wincertstore           0.2
wrapt                  1.12.1
zipp                   3.1.0

This is the error I get once i try to import tensorflow. (NOTE: I get this error on GPU install and CPU install of Tensorflow 2.1)

    Python 3.7.7 (default, May  6 2020, 11:45:54) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
2020-07-05 09:48:26.577683: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow\__init__.py", line 101, in <module>
    from tensorflow_core import *
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\__init__.py", line 46, in <module>
    from . _api.v2 import compat
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\__init__.py", line 39, in <module>
    from . import v1
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\v1\__init__.py", line 32, in <module>
    from . import compat
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\v1\compat\__init__.py", line 39, in <module>
    from . import v1
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\v1\compat\v1\__init__.py", line 29, in <module>
    from tensorflow._api.v2.compat.v1 import app
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\__init__.py", line 39, in <module>
    from . import v1
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\v1\__init__.py", line 32, in <module>
    from . import compat
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\v1\compat\__init__.py", line 39, in <module>
    from . import v1
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_core\_api\v2\compat\v1\compat\v1\__init__.py", line 667, in <module>
    from tensorflow_estimator.python.estimator.api._v1 import estimator
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_estimator\__init__.py", line 10, in <module>
    from tensorflow_estimator._api.v1 import estimator
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_estimator\_api\v1\estimator\__init__.py", line 10, in <module>
    from tensorflow_estimator._api.v1.estimator import experimental
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_estimator\_api\v1\estimator\experimental\__init__.py", line 10, in <module>
    from tensorflow_estimator.python.estimator.canned.dnn import dnn_logit_fn_builder
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_estimator\python\estimator\canned\dnn.py", line 33, in <module>
    from tensorflow_estimator.python.estimator import estimator
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_estimator\python\estimator\estimator.py", line 53, in <module>
    from tensorflow_estimator.python.estimator import util as estimator_util
  File "G:\ProgramFiles\Anaconda37\envs\tensorflow\lib\site-packages\tensorflow_estimator\python\estimator\util.py", line 75, in <module>
    class _DatasetInitializerHook(tf.compat.v1.train.SessionRunHook):
AttributeError: module 'tensorflow' has no attribute 'compat'

Answer 1

Special note for Conda users:

When one enters conda install tensorflow it installs 2.1.0 but it brings with it tensorflow-estimator 2.2.0. To fix this problem simply run conda install tensorflow-estimator==2.1.0 after installing tensorflow 2.1.0 in Conda.

This advice is valid until conda switches to TF 2.2.0 (or better yet to 2.3.0)

credits to this TF github thread

UPDATE: as of 20-Aug-2021 Conda has TF 2.4 for Windows and 2.5 for Linux. Check here to see current state of TF support in Conda.

Answer 2

the specs you have given are correct

tensorflow-gpu 2.1

python 3.7.7

CUDA 10.1

Anaconda 3.7

but in tensorflow 2.X first install

pip install tensorflow-estimator==2.1.*

then install

pip install tensorflow-gpu 

and then run the following command to check the number of gpu

import tensorflow as tf
print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))

Answer 3

This is usually caused by the broken TensorFlow-estimator module.

simply do a

pip install tensorflow-estimator==2.1.*