10

I tried installing tensorflow on my system, but I received the following error:

tensorflow-0.5.0-cp27-none-linux_x86_64.whl is not a supported wheel on this platform

Exception Information captured in pip.log file:-


/home/gansai/tensorflow/bin/pip run on Wed Nov 11 00:19:05 2015

tensorflow-0.5.0-cp27-none-linux_x86_64.whl is not a supported wheel on this platform.

Exception information: Traceback (most recent call last):

File "/home/gansai/tensorflow/local/lib/python2.7/site-packages/pip/basecommand.py", line 122, in main status = self.run(options, args)

File "/home/gansai/tensorflow/local/lib/python2.7/site-packages/pip/commands/install.py", line 269, in run InstallRequirement.from_line(name, None))

File "/home/gansai/tensorflow/local/lib/python2.7/site-packages/pip/req.py", line 168, in from_line raise UnsupportedWheel("%s is not a supported wheel on this platform." % wheel.filename)

UnsupportedWheel: tensorflow-0.5.0-cp27-none-linux_x86_64.whl is not a supported wheel on this platform.


What could I do to install tensorflow and start experimenting it?

a3.14_Infinity
  • 5,653
  • 7
  • 42
  • 66
  • The whl files are specific to a set of hardware and software What are you using? Intel processor? 32-bit or 64-bit processor? OS? etc. – Guy Coder Nov 20 '15 at 15:58
  • Intel Core 2 Duo, Debian OS - i686 (arch) – a3.14_Infinity Nov 21 '15 at 05:45
  • I have installed TensorFlow three different ways: 1 from binary on VMWare running Debian, 2 from source on VMWare running Debian, 3 from Docker on Windows. I was able to get each one to install. I did have some problems but by carefully reading the instructions I achieved success. I don't know what more to ask. Try doing a source build instead of binary install. – Guy Coder Nov 21 '15 at 13:09
  • I was using a 32 bit, but to start using TensorFlow, I installed a 64 bit debian and tried the steps mentioned by @Sasidhar below, but that also returned few errors, finally I tried docker based installation, which worked fine and started doing some examples, will post answer below. Thanks. – a3.14_Infinity Nov 22 '15 at 16:03
  • I did a Docker install but ran out of memory. Since I had TensorFlow working from the source I stuck with that. I would ask that you try a few examples before posting your answer, you may run into issues you don't forsee. Glad you have it working. – Guy Coder Nov 22 '15 at 16:06
  • Hi @GuyCoder, thanks.. I tried installing with pip again, found that I had to install python-dev package too. After that, thanks to all, I could install tensorflow and could try the examples well. So, shall I update answer with these details? – a3.14_Infinity Nov 22 '15 at 16:33

3 Answers3

10

I guess pip3 is being used for installation
it can be solved by using pip2.7

I followed the steps in here

hope it helps you:)

Community
  • 1
  • 1
Sasidhar Boddeti
  • 1,214
  • 1
  • 11
  • 22
7

tensorflow-0.5.0-cp27-none-linux_x86_64.whl is not a supported wheel on this platform

The above error comes because of trying to install TensorFlow onto a 32 bit system. As you could observe, the wheel was linux_x86_64, which is intended to be installed on 64 bit.

Steps to follow:-

Direct Binary Installation

  1. Install Debian 64 bit OS. Download debian 64 bit ISO from torrent (http://cdimage.debian.org/debian-cd/current-live/amd64/bt-hybrid/)
  2. Install python-dev. apt-get install python-dev
  3. Install tensor-flow python2.7 -m pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.5.0-cp27-none-linux_x86_64.whl

Docker Based Installation

  1. Install Debian 64 bit OS. Download debian 64 bit ISO from torrent (http://cdimage.debian.org/debian-cd/current-live/amd64/bt-hybrid/)
  2. Install Docker on Debian. Follow steps mentioned in http://docs.docker.com/engine/installation/debian/#debian-jessie-80-64-bit
  3. Run docker container for tensorflow. docker run -it b.gcr.io/tensorflow/tensorflow. Follow steps mentioned in http://tensorflow.org/get_started/os_setup.md#docker-based_installation

The main point to be noted here is that, as of now, python wheel for tensorflow is supported for 64 bit system, as shared in pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.5.0-cp27-none-linux_x86_64.whl

a3.14_Infinity
  • 5,653
  • 7
  • 42
  • 66
1

It might be unrelated but I had the same error with tensorflow-0.7.1-cp34-none-linux_x86_64.whl on a fresh Ubuntu Linux 14.04 LTS (64bit) and this is what has helped:

sudo apt-get install python3-setuptools -y && sudo easy_install3 pip -y && sudo apt-get install python 3.5-dev -y && sudo apt-get install python3.4-dev -y
sudo pip3 install --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.7.1-cp34-none-linux_x86_64.whl

You can call a tensorflow script like this:

#call it with python3.4
python3.4 tensorflow_demo.py

While researching I often read about using wget and renaming the file but this became obsolete with the latest version, see here: https://github.com/tensorflow/tensorflow/issues/1142#issuecomment-186740120

CodingYourLife
  • 7,172
  • 5
  • 55
  • 69