Switching from tensorflow to python 3.6 to python 3.5
This is my first question on stackoverflow, please bear with me as I do my best to provide as much information as possible.
I have a windowed, 6-bit processor. My ultimate goal is to use kera in the spy. The first thing I did was upgrade python to 3.6 and install tensorflow, which seemed to work. However, when I tried to get keras it didn't work and I read that keras worked on python 3.5. I have successfully installed keras on python 3.5 which automatically set theano as the backend.
But now I have two spyder environments, one working with python 3.5, one from 3.6. 3.5 reads keras but doesn't go through any modules because it can't find tensorflow. 3.6 can read tensorflow but can't find keras.
Please let me know what you would recommend. Thank!
source to share
Create a virtual disk with python 3.5 installed.
I have covered this same issue using Jupyter Notebook. Didn't understand why you would need a virtual player until I found out about this obstacle.
Detailed information on installing and configuring virtualenv can be found here: http://pymote.readthedocs.io/en/latest/install/windows_virtualenv.html
source to share
Odd, the installation instructions say TF only supports Python 3.5 on Windows. I would uninstall TF using pip uninstall tensorflow
(if you installed it using pip
) using pip
from your Python 3.6 path, then reinstall ( pip install --upgrade tensorflow
), make sure you are using pip
from your Python 3.5 path.
source to share
I had some problems with my tensorflow setup. I have personally used anaconda to solve the problem. After installing anaconda (perhaps uninstall the old one if you already have one), run the anaconda prompt and type conda create -n tensorflow python=3.5
, especially since you have to activate it with activate tensorflow
. Once this is done, you must install fiforflow on your python 3.5. To do this, use:
pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/windows/cpu/tensorflow-1.2.0rc1-cp35-cp35m-win_amd64.whl
for cpu version
pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/windows/gpu/tensorflow_gpu-1.2.0rc1-cp35-cp35m-win_amd64.whl
for gpu version
You now have tensorflow version r1.2.
Then just use pip install keras
and keras will be installed.
Now all you have to do is launch the anaconda navigator, select tensorflow from the scroll menu and launch spyder / jupyter. Now you can use Keras with tensorflow core component in Python 3.5
Hope this helped someone! (It takes me so long to find it on my own)
source to share