How do I install Keras and Theano in Anaconda Python on Windows?

问题

I am trying to work on neural networks in Python using the following Keras packages:

from keras.utils import np_utils
from keras.layers.core import Dense, Activation, Dropout
from keras.models import Sequential
from keras.layers.core import Dense, Dropout, Activation, Flatten
from keras.layers.convolutional import Convolution2D, MaxPooling2D
from keras.optimizers import SGD

But, I am getting the following error:

 15 import theano
 ---> 16 from theano import gof
 17 from theano.compat.python2x import partial
 18 import theano.compile.mode
 ImportError: cannot import name gof

Installing installed conda install keras. Later I tried to use pip install Theano, but it did not work. I Tried to install using pip install git, but I am getting this error: cannot find command git. So I installed Git and I set the environment variables.

So, is there any procedure to install these packages?

回答1:

It is my solution for the same problem

Install TDM GCC x64.
Install Anaconda x64.
Open the Anaconda prompt
Run conda update conda
Run conda update --all
Run conda install mingw libpython
Install the latest version of Theano, pip install git+git://github.com/Theano/Theano.git
Run pip install git+git://github.com/fchollet/keras.git

回答2:

The trick is that you need to create an environment/workspace for Python. This solution should work for Python 2.7 but at the time of writing keras can run on python 3.5, especially if you have the latest anaconda installed (this took me awhile to figure out so I'll outline the steps I took to install KERAS in python 3.5):

Create environment/workspace for Python 3.5

C:\conda create --name neuralnets python=3.5
C:\activate neuralnets

Install everything (notice the neuralnets workspace in parenthesis on each line). Accept any dependencies each of those steps wants to install:

(neuralnets) C:\conda install theano
(neuralnets) C:\conda install mingw libpython
(neuralnets) C:\pip install tensorflow
(neuralnets) C:\pip install keras

Test it out:

(neuralnets) C:\python -c "from keras import backend; print(backend._BACKEND)"

Just remember, if you want to work in the workspace you always have to do:

C:\activate neuralnets

so you can launch Jupyter for example (assuming you also have Jupyter installed in this environment/workspace) as:

C:\activate neuralnets
(neuralnets) jupyter notebook

You can read more about managing and creating conda environments/workspaces at the follwing URL: https://conda.io/docs/using/envs.html

回答3:

In windows with anaconda, just go on conda prompt and use this command

conda install --channel https://conda.anaconda.org/conda-forge keras

回答4:

I use macOS and used to have the same problem.
Running the following command in the terminal saved me:

conda install -c conda-forge keras tensorflow

Hope it helps.

回答5:

In case you want to train CNN's with the theano backend like the Keras mnist_cnn.py example:

You better use theano bleeding edge version. Otherwise there may occur assertion errors.

Run Theano bleeding edge
pip install --upgrade --no-deps git+git://github.com/Theano/Theano.git
Run Keras (like 1.0.8 works fine)
pip install git+git://github.com/fchollet/keras.git

回答6:

install by this command given below conda install -c conda-forge keras

this is error "CondaError: Cannot link a source that does not exist" ive get in win 10. for your error put this command in your command line.

conda update conda

this work for me .

回答7:

In windows environment with Anconda. Go to anconda prompt from start. Then if you are behind proxy then .copndarc file needs to eb updated with the proxy details.

ssl_verify: false channels: - defaults proxy_servers: http: http://xx.xx.xx.xx:xxxx https: https://xx.xx.xx.xx:xxxx

I had ssl_verify initially marked as 'True' then I was getting ssl error. So i turned it to false as above and then ran the below commands

conda update conda conda update --all conda install --channel https://conda.anaconda.org/conda-forge keras conda install --channel https://conda.anaconda.org/conda-forge tensorflow

My python version is 3.6.7