Theano used as default backend with Keras

Question

I have just installed Tensorflow and Keras in a conda environment (with miniconda).

In the keras.json file, the backend is tensorflow and there is no environment variable KERAS_BACKEND defined, however, when I import keras backend, I am told that Theano is used !

I don't know what's wrong....

Could it be related [to this answer](https://stackoverflow.com/a/43154999/567620)? — ely, Apr 18 '18 at 18:58
that's what I thought, but there is no `conda` folder in my `ENV/etc` directory — GuillaumeA, Apr 18 '18 at 19:01
clearly, conda sets environment variables when I source the virtual environment, but I don't find where.... — GuillaumeA, Apr 18 '18 at 19:07
A workaround is to prevent the environement variable loading in the `__init__.py` file of `keras.backend` so that it uses only the `keras.json` config file — GuillaumeA, Apr 18 '18 at 20:02
Is [this discussion](https://github.com/keras-team/keras/issues/6925) helpful? — Y. Luo, Apr 18 '18 at 20:43
thanks, but as said, I don't have any conda folder in my `etc` directory... — GuillaumeA, Apr 18 '18 at 20:59

GuillaumeA · Answer 1 · 2018-04-24T19:00:17.527

0

Two solutions in fact:

1- When installing Keras with conda install keras, the solution I found is to overwrite the __init.py__ file of the keras.backend module. In the file, I simply commented the portion of code where it checks for an environment variable. However, I have not found where the environment variable was set.

2- The proposed solution in other posts like this one is the solution to find where the environment variable is set. For this, keras has to be installed with the channel of conda-forge conda install -c conda-forge keras.

edited Apr 24 '18 at 19:00

answered Apr 19 '18 at 12:56

GuillaumeA

3,493
4
35
69

As pointed out by keras developpers, installing Keras with conda is not the recommanded way, only pip is supported – GuillaumeA Apr 24 '18 at 19:01

Theano used as default backend with Keras

1 Answers1