Reputation: 103
After installing tensorflow-gpu using pip3, I am getting the following error when trying to import tensorflow as tf
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/dlpda/.local/lib/python3.5/site-packages/tensorflow/__init__.py", line 22, in <module>
from tensorflow.python import pywrap_tensorflow # pylint: disable=unused-import
File "/home/dlpda/.local/lib/python3.5/site-packages/tensorflow/python/__init__.py", line 52, in <module>
from tensorflow.core.framework.graph_pb2 import *
ImportError: No module named 'tensorflow.core'
Although I am able to import successfully in a virtualenv, but not outside.
How should I resolve this problem to be able to import tensorflow?
Upvotes: 9
Views: 16996
Reputation: 1
For my issue (Error importing: No module named 'tensorflow.core.kernels.boosted_trees'
) this worked out:
pip3 install --force-reinstall tensorflow
Upvotes: 0
Reputation: 453
I was facing the same issue in tensorflow 1.15 with django 2.1.0 and python 3.7.4
It was working fine with tensorflow 1.14 with django 2.1.0 and python 3.7.4.
There is some issue in latest version of tensorflow(2.0/1.15) with previous version of django. same issue is observed with previous version of flask also.
https://github.com/tensorflow/tensorflow/issues/34607
Proposed solution is to update your django version from 2.1.0 > 2.2.5 , it works for me.
Upvotes: 0
Reputation: 9885
This is probably a bug in tensorflow. In 1.13, we have an __init__.py
that contains:
__all__ = [_s for _s in dir() if not _s.startswith('_')]
# ... later ...
try:
del python
del core
So while we remove python
and core
from the module, we do not remove them from __all__
. This prevents use cases like from tensorflow import *
. This should be fixed in Tensorflow 1.14 with this change.
Upvotes: 3
Reputation: 11
Please create a virtual environment and then install tensorflow into that. It works fine.
Upvotes: 1
Reputation: 51
The problem may be with packages installation directories like some packages are installed in home and some in /usr/. I suggest you to remove all the packages in home directory by finding them in ~/.local/lib/python3.5/site-packages and reinstall then with super user privileges.
Upvotes: 2