how to install older versions of tensor flow

Question

I am getting headache with tensorflow installations...

I have CUDA 8, CUdnn 6 and UBUNTU 16.04, python 2.7

I want now to install yensorflow. I have followed what they have explained for installation on the website but I have version ismatches of libraries. As I am using caffe, I do not want to change those versions. So I have to go for older versions of tensorflow it seems.

But I can not find a good way to install it. Every time it is just failuare. I am very tired now.

I go like this: 1- download tensorflow r1.4 from here: https://github.com/tensorflow/tensorflow/tree/r1.4

2- My 'LD_LIBRARY_PATH' is:

    :/usr/local/cuda-8.0/lib64/:/usr/local/cuda/lib64:/home/aa/torch/install/lib:/usr/lib/

3- The I run successfully this:

 $ sudo apt-get install python-pip python-dev python-virtualenv # for Python 2.7

4- and this:

$ virtualenv --system-site-packages targetDirectory # for Python 2.7

5- I activate the environment

$ source ~/tensorflow/bin/activate

6- and this

(tensorflow)$ easy_install -U pip

Till here, everything is fine.

7- and when I run this line

pip install --upgrade tensorflow-gpu

It finished successfully with this message:

Successfully installed absl-py-0.1.10 numpy-1.14.1 protobuf-3.5.1 setuptools-38.5.1 six-1.11.0 tensorflow-gpu-1.5.0 tensorflow-tensorboard-1.5.1 werkzeug-0.14.1 wheel-0.30.0

But then I cannot import it in python and this message is given:

 Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/aa/tensorflow/local/lib/python2.7/site-packages/tensorflow/__init__.py", line 24, in <module>
from tensorflow.python import *
  File "/home/aa/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/__init__.py", line 49, in <module>
from tensorflow.python import pywrap_tensorflow
   File "/home/aa/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in <module>
raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/home/aa/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
from tensorflow.python.pywrap_tensorflow_internal import *
  File "/home/aa/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
_pywrap_tensorflow_internal = swig_import_helper()
  File "/home/aa/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
_mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory


Failed to load the native TensorFlow runtime.

See https://www.tensorflow.org/install/install_sources#common_installation_problems

for some common reasons and solutions.  Include the entire stack trace above this error message when asking for help.

I think I should go for older versions.

I did install older versions but then when I test installation in python I see this messages:

>>> import tensorflow
  Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/__init__.py", line 24, in <module>
from tensorflow.python import *
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/python/__init__.py", line 52, in <module>
from tensorflow.core.framework.graph_pb2 import *
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/core/framework/graph_pb2.py", line 16, in <module>
from tensorflow.core.framework import node_def_pb2 as tensorflow_dot_core_dot_framework_dot_node__def__pb2
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/core/framework/node_def_pb2.py", line 16, in <module>
from tensorflow.core.framework import attr_value_pb2 as tensorflow_dot_core_dot_framework_dot_attr__value__pb2
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/core/framework/attr_value_pb2.py", line 16, in <module>
from tensorflow.core.framework import tensor_pb2 as tensorflow_dot_core_dot_framework_dot_tensor__pb2
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/core/framework/tensor_pb2.py", line 16, in <module>
from tensorflow.core.framework import resource_handle_pb2 as tensorflow_dot_core_dot_framework_dot_resource__handle__pb2
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/core/framework/resource_handle_pb2.py", line 22, in <module>
serialized_pb=_b('\n/tensorflow/core/framework/resource_handle.proto\x12\ntensorflow\"r\n\x13ResourceHandleProto\x12\x0e\n\x06\x64\x65vice\x18\x01 \x01(\t\x12\x11\n\tcontainer\x18\x02 \x01(\t\x12\x0c\n\x04name\x18\x03 \x01(\t\x12\x11\n\thash_code\x18\x04 \x01(\x04\x12\x17\n\x0fmaybe_type_name\x18\x05 \x01(\tB/\n\x18org.tensorflow.frameworkB\x0eResourceHandleP\x01\xf8\x01\x01\x62\x06proto3')
TypeError: __init__() got an unexpected keyword argument 'syntax'


>>> import tensorflow
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/__init__.py", line 24, in <module>
from tensorflow.python import *
  File "/home/aa/.local/lib/python2.7/site-packages/tensorflow/python/__init__.py", line 49, in <module>
from tensorflow.python import pywrap_tensorflow
ImportError: cannot import name pywrap_tensorflow

What does it say when you try to import TensorFlow to python? — Pam
– Pam, Commented Feb 22, 2018 at 18:53
And LD_LIBRARY_PATH ends with "No such file or directory". Why’s that? — Pam
– Pam, Commented Feb 22, 2018 at 18:58
I put the error message above. Answer to the next comment: yes, I have that folder and I have 3 files in it: 'libcupti.so','libcupti.so.8.0' and 'libcupti.so.8.0.61'. I think that is due to extra (:)! I also do not know how to fix it:( — azzz
– azzz, Commented Feb 22, 2018 at 19:17
I tried to remove the end by this: 'export LD_LIBRARY_PATH="${LD_LIBRARY_PATH/:${:/usr/local/cuda/extras/CUPTI/lib64}/}" ' but it gave me an error like: 'bash: :${:/usr/local/cuda/extras/CUPTI/lib64}: bad substitution' — azzz
– azzz, Commented Feb 22, 2018 at 19:18
So the error looks like it doesn’t find libcublas.so (it's a library file). Can you find that? And your comment above, could you use: export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/mypath I don’t think you need all the {}. — Pam
– Pam, Commented Feb 22, 2018 at 19:30

Nilav Baran Ghosh · Accepted Answer · 2018-02-22 19:18:54Z

4

You should be able to install earlier versions of tensorflow using

pip install tensorflow==1.3 where 1.3 is the version number. You can pick the version of your choice.

answered Feb 22, 2018 at 19:18

Nilav Baran Ghosh

1,34911 silver badges18 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

azzz Over a year ago

I did it but still I get error as I have written at the end of my message. I did install version 1.3.0 and 1.4.1 but none worked.

Pam Over a year ago

And TensorFlow-gpu?

Nilav Baran Ghosh Over a year ago

pip install tensorflow-gpu==1.3 should give you the gpu enabled library

Supun Abesekara · Accepted Answer · 2018-02-23 04:00:10Z

3

you can find .whl file of tensorflow in here https://pypi.python.org/pypi/tensorflow/1.4.0 and install it using pip. hope you looking tensorflow 1.4.0 version in case if you want older version change url https://pypi.python.org/pypi/tensorflow/1.x.0(x= version you want)

edited Feb 23, 2018 at 4:00

Supun Abesekara

7368 silver badges33 bronze badges

answered Feb 22, 2018 at 19:28

Chathuranga Kalana

3434 silver badges12 bronze badges

Comments

azzz · Accepted Answer · 2018-03-01 16:31:47Z

0

'The problem is solved. I had "/usr/lib/python2.7/dist-packages" in my PYTHONPATH. So there was a version mismatch between protobufs and it was solved by typing 'unset PYTHONPATH' before starting to import tensorflow.

answered Mar 1, 2018 at 16:31

azzz

917 bronze badges

Collectives™ on Stack Overflow

how to install older versions of tensor flow

3 Answers 3

3 Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related