Reputation: 515
I've installed both anaconda 2 and 3 in a Windows 64bit machine. I'm trying to install Tensorflow with anaconda as per the instructions here. However, I'm getting the following error while doing a pip install.
tensorflow-1.0.0-cp35-cp35m-win_x86_64.whl is not a supported wheel on this platform.
My anaconda info:
platform : win-64
conda version : 4.2.13
conda is private : False
conda-env version : 4.2.13
conda-build version : 1.20.0
python version : 2.7.12.final.0
requests version : 2.13.0
root environment : C:\Anaconda2 (writable)
default environment : C:\Anaconda2\envs\tensorflow
envs directories : C:\Anaconda2\envs
package cache : C:\Anaconda2\pkgs
channel URLs : https://repo.continuum.io/pkgs/free/win-64
https://repo.continuum.io/pkgs/free/noarch
https://repo.continuum.io/pkgs/pro/win-64
https://repo.continuum.io/pkgs/pro/noarch
https://repo.continuum.io/pkgs/msys2/win-64
https://repo.continuum.io/pkgs/msys2/noarch
config file : None
offline mode : False
I would really appreciate any help since I'm banging my head on this since last 8 hours.
I've removed Anaconda 2 and tried to install tensorflow again. Getting the same error. Please find the anaconda info.
platform : win-64
conda version : 4.2.9
conda is private : False
conda-env version : 4.2.9
conda-build version : 2.0.2
python version : 3.5.2.final.0
requests version : 2.11.1
root environment : C:\Users\Gramener\Anaconda3 (writable)
default environment : C:\Users\Gramener\Anaconda3\envs\tensorflow
envs directories : C:\Users\Gramener\Anaconda3\envs
package cache : C:\Users\Gramener\Anaconda3\pkgs
channel URLs : https://repo.continuum.io/pkgs/free/win-64/
https://repo.continuum.io/pkgs/free/noarch/
https://repo.continuum.io/pkgs/pro/win-64/
https://repo.continuum.io/pkgs/pro/noarch/
https://repo.continuum.io/pkgs/msys2/win-64/
https://repo.continuum.io/pkgs/msys2/noarch/
config file : None
offline mode : False
Upvotes: 7
Views: 14211
Reputation: 11
Try this:
pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/windows/cpu/tensorflow-1.0.1-cp36-cp36m-win_amd64.whl
It is the command that runs when you install with native pip. This change worked out for me
Upvotes: 0
Reputation: 4065
Explicitly use python 3.5 when you create your Anaconda environment. IE:
create -n tensorflow python=3.5
then run the following command to install tensorflow:
pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/windows/cpu/tensorflow-1.0.1-cp35-cp35m-win_amd64.whl
Then you can successful do the Hello Tensorflow example from https://www.tensorflow.org/install/install_windows
Upvotes: 9
Reputation: 83197
On Microsoft Windows, TensorFlow needs Python 3.5 64-bit. You seem to use Python 2.7.
Upvotes: 0
Reputation: 141
I had the same problem, I uninstalled Anaconda and installed python 3.5.2 from python.org and I was able to run Hello Tensorflow.
Upvotes: 0
Reputation:
Believe me I have survived it.
First, please ensure that your system is Win7 64Bit and your Python version is 3.5.2(64bit), I didn't try Python 3.5.
Second, don't believe Anaconda and PyCharm. Of course I know they can both be used to install packages. However, the packages installed by them can't be synchronized to cmd sometimes.
what I want to say is: Ensure the content of cmd is the same as the website of TensorFlow.
I wanted to upload more pictures to tell you how to do, but stackoverflow didn't allow it.
Details are shown as follows:
Upvotes: 1
Reputation: 47
I've encountered the same problem recently when trying to install tensorflow v1.0 on azure vm. None of anaconda versions worked for me.
So I decided to try upgrading tensorflow to v1.0 on my local pc to see if the error will be reproducible. This resulted in the same error (even though I've installed r0.12 successfully just couple weeks ago).
Reading tensorflow installation notes carefully I've noticed the following sentence:
The Anaconda installation is community supported, not officially supported.
So I've decided to try python 3.5.2 from python.org and this worked immediately.
I believe the problem here is that the latest tensorflow version (1.0) does not work with anaconda yet. Nevertheless if you still want to use anaconda you can try tensorflow r0.12, but keep in mind that doing so is not recommended and there were breaking changes from r0.12 to 1.0
Upvotes: 2