How to reduce tensor flow, maybe several versions?

Question

I have tensorflow 1.2.1 installed and I need to downgrade it to version 1.1 in order to run a specific tutorial. What is the safe way to do this? I am using Windows 10, python 3.5. Tensorflow was installed with pip3, but "pip3 show tensorflow" returns empty.

Is it possible to have several versions of tensor flow on the same OS?

Answer 1

Pip allows you to specify the version

pip install tensorflow==1.1

Answer 2

Can I have multiple versions of tenorflow on the same OS?

Yes, you can use Python virtual environments to do this. From the docs :

A virtual environment is a tool for storing the dependencies needed by various projects in different places, creating virtual Python environments for them. It solves the problem "Project X depends on version 1.x, but Project Y requires 4.x" and keeps your global site-packages directory clean and manageable.

After you installed virtualenv (see the Documentation ), you can create a virtual environment for the tutorial and install the version of tenorflow you need in it:

 PATH_TO_PYTHON=/usr/bin/python3.5 virtualenv -p $PATH_TO_PYTHON my_tutorial_env source my_tutorial_env/bin/activate # this activates your new environment pip install tensorflow==1.1 

PATH_TO_PYTHON should indicate where Python is installed on your system. When you want to use another version of tenorflow, do:

 deactivate my_tutorial_env 

Now you can work again with the version of tenorflow that has already been installed on your system.

Answer 3

I discovered the joy of the anaconda: https://www.continuum.io/downloads

 C:> conda create -n tensorflow1.1 python=3.5 C:> activate tensorflow1.1 (tensorflow1.1) C:> pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/windows/gpu/tensorflow_gpu-1.1.0-cp35-cp35m-win_amd64.whl 

voila creates a virtual environment.

Answer 4

If you use python3 for windows, you can also do this

pip3 install tensorflow==1.4

you can choose any version from "(from versions: 1.2.0rc2, 1.2.0, 1.2.1, 1.3.0rc0, 1.3.0rc1, 1.3.0rc2, 1.3.0, 1.4.0rc0, 1.4.0rc1, 1.4.0 , 1.5.0rc0, 1.5.0rc1, 1.5.0, 1.5.1, 1.6.0rc0, 1.6.0rc1, 1.6.0, 1.7.0rc0, 1.7.0rc1, 1.7.0) "

I did this when I wanted to downgrade from 1.7 to 1.4

Answer 5

You can try using the --no-cache-dir options along with -I to overwrite the cache of the previous version and install the new version. For instance:

 pip3 install --no-cache-dir -I tensorflow==1.1 

Then use the following command to check the version of tenorflow:

 python3 -c 'import tensorflow as tf; print(tf.__version__) 

This should show the correct version that was installed.