[python] TensorFlow not found using pip

I'm trying to install TensorFlow using pip:

$ pip install tensorflow --user
Collecting tensorflow
Could not find a version that satisfies the requirement tensorflow (from versions: )
No matching distribution found for tensorflow

What am I doing wrong? So far I've used Python and pip with no issues.

The above answers helped me to solve my issue specially the first answer. But adding to that point after the checking the version of python and we need it to be 64 bit version.

Based on the operating system you have we can use the following command to install tensorflow using pip command.

The following link has google api links which can be added at the end of the following command to install tensorflow in your respective machine.

Root command: python -m pip install --upgrade (link) link : respective OS link present in this link

If you are trying to install Tensorflow with Anaconda on Windows, a free advice is to please uninstall anaconda and download a 64-bit Python version, ending with amd64 from releases page. For me, its python-3.7.8-amd64.exe

Then install Tensorflow in a virtual environment by following the instructions on official website of Tensorflow.

Install Python 3.5.x 64 bit amd version here. Make sure you add Python to your PATH variable. Then open a command prompt and type

python -m pip install --upgrade pip

should give you the following result :

 Collecting pip
 Using cached pip-9.0.1-py2.py3-none-any.whl
 Installing collected packages: pip
 Found existing installation: pip 7.1.2
 Uninstalling pip-7.1.2:
 Successfully uninstalled pip-7.1.2
 Successfully installed pip-9.0.1

Now type

 pip3 install --upgrade tensorflow

Currently PIP does not have a 32bit version of tensorflow, it worked when I uninstalled python 32bit and installed x64

This worked for me with Python 2.7 on Mac OS X Yosemite 10.10.5:

sudo pip install --upgrade https://storage.googleapis.com/tensorflow/mac/tensorflow-0.5.0-py2-none-any.whl

1. Install python by checking Add Python to Path
2. pip3 install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl

This works for windows 10.0

I installed tensorflow on conda but didnt seem to work on windows but finally this command here works fine on cmd.

 python.exe -m pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl

Updated 11/28/2016: TensorFlow is now available in PyPI, starting with release 0.12. You can type

pip install tensorflow

...or...

pip install tensorflow-gpu

...to install the CPU-only or GPU-accelerated version of TensorFlow respectively.

Previous answer: TensorFlow is not yet in the PyPI repository, so you have to specify the URL to the appropriate "wheel file" for your operating system and Python version.

The full list of supported configurations is listed on the TensorFlow website, but for example, to install version 0.10 for Python 2.7 on Linux, using CPU only, you would type the following command:

$ pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.10.0rc0-cp27-none-linux_x86_64.whl

The correct way to install it would be as mentioned here

$ pip install --upgrade TF_BINARY_URL   # Python 2.7
$ pip3 install --upgrade TF_BINARY_URL  # Python 3.N

Find the correct TF_BINARY_URL for your hardware from the tensor flow official homepage

Following these steps allows you to install tensorflow and keras:

1. Download Anaconda3-5.2.0 which comes with python 3.6 from https://repo.anaconda.com/archive/

2. Install Anaconda and open Anaconda Prompt and execute below commands

   conda install jupyter 
   conda install scipy
   pip install sklearn
   pip install msgpack
   pip install pandas
   pip install pandas-datareader
   pip install matplotlib 
   pip install pillow
   pip install requests
   pip install h5py
   pip install tensorflow
   pip install keras
   

If you are using the Anaconda Python installation, pip install tensorflow will give the error stated above, shown below:

Collecting tensorflow
Could not find a version that satisfies the requirement tensorflow (from versions: )
No matching distribution found for tensorflow

According to the TensorFlow installation page, you will need to use the --ignore-installed flag when running pip install.

However, before this can be done see this link to ensure the TF_BINARY_URL variable is set correctly in relation to the desired version of TensorFlow that you wish to install.

1. Start Command Prompt with Administrative Permission
2. Enter following command python -m pip install --upgrade pip
3. Next Enter command pip install tensorflow

For Window you can use below command

python3 -m pip install --upgrade https://storage.googleapis.com/tensorflow/windows/gpu/tensorflow_gpu-2.3.0-cp38-cp38-win_amd64.whl

My env: Win 10, python 3.6

pip3 install --upgrade tensorflow
pip install --upgrade tensorflow

With error:

> Collecting tensorflow Could not find a version that satisfies the
> requirement tensorflow (from versions: ) No matching distribution
> found for tensorflow

I also tried pip install tensorflow and pip install tensorflow-gpu. But error:

> Could not find a version that satisfies the requirement tensorflow (from versions: ) No matching distribution found for tensorflow
> Could not find a version that satisfies the requirement tensorflow-gpu (from versions: ) No matching distribution found for tensorflow-gpu

Install OK when tried with Step: (https://www.tensorflow.org/install/install_windows)

1. Follow the instructions on the Anaconda download site to download and install Anaconda. https://www.continuum.io/downloads

2. Create a conda environment named tensorflow by invoking the following command:

   C:> conda create -n tensorflow pip python=3.5 
   

3. Activate the conda environment by issuing the following command:

   C:> activate tensorflow
    (tensorflow)C:>  # Your prompt should change 
   

4. Issue the appropriate command to install TensorFlow inside your conda environment. To install the CPU-only version of TensorFlow, enter the following command:

   (tensorflow)C:> pip install --ignore-installed --upgrade tensorflow 
   

   To install the GPU version of TensorFlow, enter the following command (on a single line):

   (tensorflow)C:> pip install --ignore-installed --upgrade tensorflow-gpu 
   

From tensorflow website: "You will need pip version 8.1 or later for the following commands to work". Run this command to upgrade your pip, then try install tensorflow again:

pip install --upgrade pip

I had the same error when trying to install on my Mac (using Python 2.7). A similar solution to the one I'm giving here also seemed to work for Python 3 on Windows 8.1 according to a different answer on this page by Yash Kumar Verma

Solution

Step 1: go to The URL of the TensorFlow Python package section of the TensorFlow installation page and copy the URL of the relevant link for your Python installation.

Step 2: open a terminal/command prompt and run the following command:
pip install --upgrade [paste copied url link here]

So for me it was the following:
pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-1.2.0-py2-none-any.whl

Update (July 21 2017): I tried this with some others who were running on Windows machines with Python 3.6 and they had to change the line in Step 2 to: python -m pip install [paste copied url link here]

Update (26 July 2018): For Python 3.6.2 (not 3.7 because it's in 3.6.2 in TF Documentation), you can also use pip3 install --upgrade [paste copied URL here] in Step 2.

You need a 64-bit version of Python and in your case are using a 32-bit version. As of now Tensorflow only supports 64-bit versions of Python 3.5.x and 3.8.x on Windows. See the install docs to see what is currently supported

To check which version of Python you are running, type python or python3 to start the interpreter, and then type import struct;print(struct.calcsize("P") * 8) and that will print either 32 or 64 to tell you which bit version of Python you are running.

From comments:

To download a different version of Python for Windows, go to python.org/downloads/windows and scroll down until you see the version you want that ends in a "64". That will be the 64 bit version that should work with tensorflow

I had the same problem and solved with this:

# Ubuntu/Linux 64-bit, CPU only, Python 2.7
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.12.1-cp27-none-linux_x86_64.whl

# Ubuntu/Linux 64-bit, GPU enabled, Python 2.7
# Requires CUDA toolkit 8.0 and CuDNN v5. For other versions, see "Installing from sources" below.

# Mac OS X, CPU only, Python 2.7:
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.1-py2-none-any.whl

# Mac OS X, GPU enabled, Python 2.7:
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/gpu/tensorflow_gpu-0.12.1-py2-none-any.whl

# Ubuntu/Linux 64-bit, CPU only, Python 3.4
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.12.1-cp34-cp34m-linux_x86_64.whl

# Ubuntu/Linux 64-bit, GPU enabled, Python 3.4
# Requires CUDA toolkit 8.0 and CuDNN v5. For other versions, see "Installing from sources" below.
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-0.12.1-cp34-cp34m-linux_x86_64.whl

# Ubuntu/Linux 64-bit, CPU only, Python 3.5
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.12.1-cp35-cp35m-linux_x86_64.whl

# Requires CUDA toolkit 8.0 and CuDNN v5. For other versions, see "Installing from sources" below.
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-0.12.1-cp35-cp35m-linux_x86_64.whl

# Mac OS X, CPU only, Python 3.4 or 3.5:
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.1-py3-none-any.whl

# Mac OS X, GPU enabled, Python 3.4 or 3.5:
(tensorflow)$ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/gpu/tensorflow_gpu-0.12.1-py3-none-any.whl

Plus:

# Python 2
(tensorflow)$ pip install --upgrade $TF_BINARY_URL

# Python 3
(tensorflow)$ pip3 install --upgrade $TF_BINARY_URL

Found on Docs.

UPDATE!

There are new links for new versions

For example, for installing tensorflow v1.0.0 in OSX you need to use:

https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-1.0.0-py2-none-any.whl

instead of

https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.1-py2-none-any.whl

If you are trying to install it on a windows machine you need to have a 64-bit version of python 3.5. This is the only way to actually install it. From the website:

TensorFlow supports only 64-bit Python 3.5 on Windows. We have tested the pip packages with the following distributions of Python:

Python 3.5 from Anaconda

Python 3.5 from python.org.

You can download the proper version of python from here (make sure you grab one of the ones that says "Windows x86-64")

You should now be able to install with pip install tensorflow or python -m pip install tensorflow (make sure that you are using the right pip, from python3, if you have both python2 and python3 installed)

Remember to install Anaconda 3-5.2.0 as the latest version which is 3-5.3.0 have python version 3.7 which is not supported by Tensorflow.

Tensorflow DOES NOT support python versions after 3.8 as of when I'm writing this at least (December 2020). Use this: https://www.tensorflow.org/install to check what python versions it supports, I just spent hours looking through these answers, took me way too long to realise that.

I found this to finally work.

python3 -m pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-1.12.0-py3-none-any.whl

Edit 1: This was tested on Windows (8, 8.1, 10), Mac and Linux. Change python3 to python according to your configuration. Change py3 to py2 in the url if you are using Python 2.x.

Edit 2: A list of different versions if someone needs: https://storage.googleapis.com/tensorflow

Edit 3: A list of urls for the available wheel packages is available here: https://www.tensorflow.org/install/pip#package-location

update 2019: for install the preview version of TensorFlow 2 in Google Colab you can use:

!wget https://developer.nvidia.com/compute/cuda/10.0/Prod/local_installers/cuda-repo-ubuntu1604-10-0-local-10.0.130-410.48_1.0-1_amd64 -O cuda-repo-ubuntu1604-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
!dpkg -i cuda-repo-ubuntu1604-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
!apt-key add /var/cuda-repo-10-0-local-10.0.130-410.48/7fa2af80.pub
!apt-get update
!apt-get install cuda
!pip install tf-nightly-gpu-2.0-preview

and for install the TensorFlow 2 bye pip you can use: pip install tf-nightly-gpu-2.0-preview for GPU and pip install tf-nightly-2.0-preview for CPU.

Try this, it should work:

 python.exe -m pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl

I had this problem on OSX Sierra 10.12.2. It turns out I had the wrong version of Python installed (I had Python 3.4 but tensorflow pypi packages for OSX are only for python 3.5 and up).

The solution was to install Python 3.6. Here's what I did to get it working. Note: I used Homebrew to install Python 3.6, you could do the same by using the Python 3.6 installer on python.org

brew uninstall python3
brew install python3
python3 --version # Verify that you see "Python 3.6.0"
pip install tensorflow # With python 3.6 the install succeeds
pip install jupyter # "ipython notebook" didn't work for me until I installed jupyter
ipython notebook # Finally works!

I have experienced the same error while I tried to install tensorflow in an anaconda package. After struggling a lot, I finally found an easy way to install any package without running into an error. First create an environment in your anaconda administrator using this command

conda create -n packages

Now activate that environment

activate packages 

and try running

pip install tensorflow 

After a successful installation, we need to make this environment accessible to jupyter notebook. For that, you need to install a package called ipykernel using this command

pip install ipykernel

After installing ipykernel enter the following command

python -m ipykernel install --user --name=packages

After running this command, this environment will be added to jupyter notebook and that's it. Just go to your jupyter notebook, click on new notebook, and you can see your environment. Select that environment and try importing tensorflow and in case if you want to install any other packages, just activate the environment and install those packages and use that environment in your jupyter

Note: This answer is for Cygwin users

Leaving this answer because none of the others here worked for my use case (using the *nix-on-Windows terminal environment to install tensorflow on a virtualenv, cygwin (http://www.cygwin.com/)) (at least a simple control+F on the answer pages found nothing).

TLDR: If you are using a virtualenv in a cygwin terminal, know that cygwin seems to have a problem installing tensorflow and throws the error specified in this post's question (a similar sentiment can be found here (https://stackoverflow.com/a/45230106/8236733) (similar cause, different error)). Solved by creating the virtualenv in the Windows Command Prompt. Then can access / activate the virtualenv from a cygwin terminal via source ./Scripts/activate to use Windows' (not cygwin's) python.

When just using cygwin's python3 to try use tensorflow, eg. something like...

apt-cyg install python3-devel
cd python-virtualenv-base
virtualenv -p `which python3` tensorflow-examples

found that there were some problems with installing tensorflow-gpu package using cygwin's python. Was seeing the error

$ pip install tensorflow --user
Collecting tensorflow
Could not find a version that satisfies the requirement tensorflow (from versions: )
No matching distribution found for tensorflow

There are many proposed solutions, none of them helped in my case (they are all generally along the lines of "You probably have python3 for 32-bit achitectures installed, tensorflow requires 64-bit" or some other python mismatch mistake (whereas here, it's simply seems to be that cygwin's python had problems installing tensorflow-gpu)).

What did end up working for me was doing...

1. Install python3 via the official Windows way for the Windows system (the cygwin system is separate, so uses a different python)
2. Open the Command Prompt in Windows (not a cygwin terminal) and do...

C:\Users\me\python-virtualenvs-base>python
Python 3.6.2 (v3.6.2:5fd33b5, Jul  8 2017, 04:57:36) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> exit()

C:\Users\me\python-virtualenvs-base>pip -V
pip 9.0.1 from c:\users\me\appdata\local\programs\python\python36\lib\site-packages (python 3.6)

C:\Users\me\python-virtualenvs-base>pip install virtualenv
Collecting virtualenv
  Downloading https://files.pythonhosted.org/packages/b6/30/96a02b2287098b23b875bc8c2f58071c35d2efe84f747b64d523721dc2b5/virtualenv-16.0.0-py2.py3-none-any.whl (1.9MB)
    100% |¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦| 1.9MB 435kB/s
Installing collected packages: virtualenv
Successfully installed virtualenv-16.0.0
You are using pip version 9.0.1, however version 18.0 is available.
You should consider upgrading via the 'python -m pip install --upgrade pip' command.

C:\Users\me\python-virtualenvs-base>virtualenv tensorflow-examples
Using base prefix 'c:\\users\\me\\appdata\\local\\programs\\python\\python36'
New python executable in C:\Users\me\python-virtualenvs-base\tensorflow-examples\Scripts\python.exe
Installing setuptools, pip, wheel...done.

3. Then, can go back to the cygwin terminal, navigate back to that virtualenv that you created in the command prompt and do...

    ?  tensorflow-examples source ./Scripts/activate
    (tensorflow-examples) ?  tensorflow-examples python -V
    Python 3.6.2
    (tensorflow-examples) ?  tensorflow-examples pip install tensorflow-gpu
    Collecting tensorflow-gpu
      Downloading 
      ....

Notice you don't do source ./bin/activate in the virtualenv as you would if you had created the virtualenv in cygwin's pseudo-linux environment, but instead do source ./Scripts/activate.

I've found out the problem.

I'm using a Windows computer which has Python 2 installed previously. After Python 3 is installed (without setting the path, I successfully check the version of pip3 - but the python executable file still points to the Python2)

Then I set the path to the python3 executable file (remove all python2 paths) then start a new command prompt, try to reinstall Tensorflow. It works!

I think this problem could happend on MAC OS too since there is a default python which is on the MAC system.

I figured out that TensorFlow 1.12.0 only works with Python version 3.5.2. I had Python 3.7 but that didn't work. So, I had to downgrade Python and then I could install TensorFlow to make it work.

To downgrade your python version from 3.7 to 3.6

conda install python=3.6.8

There are multiple groups of answers to this question. This answer aims to generalize one group of answers:

There may not be a version of TensorFlow that is compatible with your version of Python. This is particularly true if you're using a new release of Python. For example, there may be a delay between the release of a new version of Python and the release of TensorFlow for that version of Python.

In this case, I believe your options are to:

1. Upgrade or downgrade to a different version of Python. (Virtual environments are good for this, e.g. conda install python=3.6)
2. Select a specific version of tensorflow that is compatible with your version of python, e.g. if you're still using python3.4: pip install tensorflow==2.0
3. Compile TensorFlow from the source code.
4. Wait for a new release of TensorFlow which is compatible with your version of Python.

I had the same problem. After uninstalling the 32-bit version of python and reinstalling the 64-bit version I tried reinstalling TensorFlow and it worked.

Link to TensorFlow guide: https://www.tensorflow.org/install/install_windows

For windows this worked for me,

Download the wheel from this link. Then from command line navigate to your download folder where the wheel is present and simply type in the following command -

pip install tensorflow-1.0.0-cp36-cp36m-win_amd64.whl

For pyCharm users:

1. Check pip version: pip3 -V
2. If pip is older than 9.0.1: py -3 -m pip install --upgrade pip
3. Then: py -3 -m pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl

The only thing that worked for me was to use Ananconda and create a new conda env with conda create -n tensorflow python=3.5 then activate using activate tensorflow and finally conda install -c conda-forge tensorflow.

This works around every issue I had including ssl certs, proxy settings, and does not need admin access. It should be noted that this is not directly supported by the tensorflow team.

Source

Try this:

export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.1-py3-none-any.whl
pip3 install --upgrade $TF_BINARY_URL

Source: https://www.tensorflow.org/get_started/os_setup (page no longer exists)

Update 2/23/17 Documentation moved to: https://www.tensorflow.org/install

If your command pip install --upgrade tensorflowcompiles, then your version of tensorflow should be the newest. I personally prefer to use anaconda. You can easily install and upgrade tensorflow as follows:

 conda install -c conda-forge tensorflow  # to install
 conda upgrade -c conda-forge tensorflow  # to upgrade

Also if you want to use it with your GPU you have an easy install:

 conda install -c anaconda tensorflow-gpu

I've been using it for a while now and I have never had any problem.

You need to use right version of Python and pip

On Windows 10, with Python 3.6.X version I was facing same then after checking deliberately , I noticed I had Python-32 bit installation on my 64 bit machine. Remember TensorFlow is only compatible with 64bit installation of python. Not 32 bit of Python

If we download Python from python.org , the default installation would be 32 bit. So we have to download 64 bit installer manually to install Python 64 bit. And then add below to PATH environment.

C:\Users\AppData\Local\Programs\Python\Python36
C:\Users\AppData\Local\Programs\Python\Python36\Scripts

Then run gpupdate /Force on command prompt. If python command doesnt work for 64 bit restart your machine.

Then run python on command prompt. It should show 64 bit

C:\Users\YOURNAME>python
Python 3.6.3 (v3.6.3:2c5fed8, Oct  3 2017, 18:11:49) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.

Then run below command to install tensorflow CPU version(recommended)

pip3 install --upgrade tensorflow

October 2020 update:

Tensorflow now supports Python 3.5.x through Python 3.8.x, but you still have to use a 64-bit version.

If you need to run multiple versions of Python on the same machine, you can use a virtual environment to help manage them.

Nothing here worked for me on Windows 10. Perhaps an updated solution below that did work for me.

python -m pip install --upgrade tensorflow.

This is using Python 3.6 and tensorflow 1.5 on Windows 10

Excerpt from tensorflow website https://www.tensorflow.org/install/install_windows

Installing with native pip

If the following version of Python is not installed on your machine, install it now:

Python 3.5.x from python.org TensorFlow only supports version 3.5.x of Python on Windows. Note that Python 3.5.x comes with the pip3 package manager, which is the program you'll use to install TensorFlow.

To install TensorFlow, start a terminal. Then issue the appropriate pip3 install command in that terminal. To install the CPU-only version of TensorFlow, enter the following command:

C:\> pip3 install --upgrade tensorflow
To install the GPU version of TensorFlow, enter the following command:

C:\> pip3 install --upgrade tensorflow-gpu

If you're trying to install tensorflow in anaconda and it isn't working, then you may need to downgrade python version because only 3.6.x is currently supported while anaconda has the latest version.

1. check python version: python --version

2. if version > 3.6.x then follow step 3, otherwise stop, the problem may be somewhere else

3. conda search python

4. conda install python=3.6.6

5. Check version again: python --version

6. If version is correct, install tensorflow (step 7)

7. pip install tensorflow

Here is what I did for Windows 10! I also did not disturb my previous installation of Python 2.7

Step1: Install Windows x86-64 executable installer from the link: https://www.python.org/downloads/release/python-352/

Step2: Open cmd as Administrator

Step3: Type this command:

pip install https://storage.googleapis.com/tensorflow/windows/cpu/tensorflow-0.12.0rc0-cp35-cp35m-win_amd64.whl

You should see that it works and as shown in the picture below, I also tried the sample example.

Unfortunately my reputation is to low to command underneath @Sujoy answer.

In their docs they claim to support python 3.6. The link provided by @mayur shows that their is indeed only a python3.5 wheel package. This is my try to install tensorflow:

Microsoft Windows [Version 10.0.16299.371]
(c) 2017 Microsoft Corporation. All rights reserved.

C:\>python3 -m pip install --upgrade pip
Requirement already up-to-date: pip in d:\python\v3\lib\site-packages (10.0.0)

C:\>python3 -m pip -V
pip 10.0.0 from D:\Python\V3\lib\site-packages\pip (python 3.6)

C:\>python3 -m pip install --upgrade tensorflow
Collecting tensorflow
Could not find a version that satisfies the requirement tensorflow (from versions: )
No matching distribution found for tensorflow

while python 3.5 seems to install successfully. I would love to see a python3.6 version since they claim it should also work on python3.6.

Quoted :

"TensorFlow supports Python 3.5.x and 3.6.x on Windows. Note that Python 3 comes with the pip3 package manager, which is the program you'll use to install TensorFlow."

Source : https://www.tensorflow.org/install/install_windows

Python3.5 install :

Microsoft Windows [Version 10.0.16299.371]
(c) 2017 Microsoft Corporation. All rights reserved.

C:\>python3 -m pip install --upgrade pip
Requirement already up-to-date: pip in d:\python\v3\lib\site-packages (10.0.0)

C:\>python3 -m pip -V
pip 10.0.0 from D:\Python\V3_5\lib\site-packages\pip (python 3.5.2)

C:\>python3 -m pip install --upgrade tensorflow
Collecting tensorflow
  Downloading 
    ....
    ....

I hope i am terrible wrong here but if not ring a alarm bell

Edit: A couple of posts below someone pointed out that the following command would work and it did.

python3 -m pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl

Strange pip is not working

2.0 COMPATIBLE SOLUTION:

Execute the below commands in Terminal (Linux/MacOS) or in Command Prompt (Windows) to install Tensorflow 2.0 using Pip:

#Install tensorflow using pip virtual env 
pip install virtualenv
virtualenv tf_2.0.0   # tf_2.0.0 is virtual env name
source tf_2.0.0/bin/activate
#You should see tf_2.0.0 Env now. Execute the below steps
pip install tensorflow==2.0.0
python
>>import tensorflow as tf
>>tf.__version__
2.0.0

Execute the below commands in Terminal (Linux/MacOS) or in Command Prompt (Windows) to install Tensorflow 2.0 using Bazel:

git clone https://github.com/tensorflow/tensorflow.git
cd tensorflow

#The repo defaults to the master development branch. You can also checkout a release branch to build:
git checkout r2.0

#Configure the Build => Use the Below line for Windows Machine
python ./configure.py 

#Configure the Build => Use the Below line for Linux/MacOS Machine
./configure
#This script prompts you for the location of TensorFlow dependencies and asks for additional build configuration options. 

#Build Tensorflow package

#CPU support
bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package 

#GPU support
bazel build --config=opt --config=cuda --define=no_tensorflow_py_deps=true //tensorflow/tools/pip_package:build_pip_package

You may try this

pip install --upgrade tensorflow

as of today, if anyone else is wondering, python >= 3.9 will cause the same issue uninstall python 3.9, and install 3.8 , it should resolve it

It seems there could be multiple reasons for tensorFlow not getting installed via pip. The one I faced on windows 10 was that I didn't had supported version of cudnn in my system path. As of now [Dec 2017], tensorflow on windows only supports cudnn v6.1. So, provide the path of cudnn 6.1, if everything else is correct then tensorflow should be installed.

Check https://pypi.python.org/pypi/tensorflow to see which packages are available.

As of this writing, they don't provide a source package, so if there's no prebuilt one for your platform, this error occurs. If you add -v to the pip command line, you'll see it iterating over the packages that are available at PyPI and discarding them for being incompatible.

You need to either find a prebuilt package somewhere else, or compile tensorflow yourself from its sources by instructions at https://www.tensorflow.org/install/install_sources .

They have a good reason for not building for some platforms though:

• A win32 package is missing because TensorFlow's dependency, Bazel, only supports win64.
• For win64, only 3.5+ is supported because earlier versions are compiled with compilers without C++11 support.

I was having this problem too. When looking at the different .whl files. I noticed there was no 32-bit version of tensorflow on python 3.7. In the end just had to install 64bit Python 3.7 from here.

Here is my Environment (Windows 10 with NVIDIA GPU). I wanted to install TensorFlow 1.12-gpu and failed multiple times but was able to solve by following the below approach.

This is to help Installing TensorFlow-GPU on Windows 10 Systems

Steps:

1. Make sure you have NVIDIA graphic card

a. Go to windows explorer, open device manager-->check “Display Adaptors”-->it will show (ex. NVIDIA GeForce) if you have GPU else it will show “HD Graphics”

b. If the GPU is AMD’s then tensorflow doesn’t support AMD’s GPU

2. If you have a GPU, check whether the GPU supports CUDA features or not.

a. If you find your GPU model at this link, then it supports CUDA.

b. If you don’t have CUDA enabled GPU, then you can install only tensorflow (without gpu)

3. Tensorflow requires python-64bit version. Uninstall any python dependencies

a. Go to control panel-->search for “Programs and Features”, and search “python”

b. Uninstall things like anaconda and any pythons related plugins. These dependencies might interfere with the tensorflow-GPU installation.

c. Make sure python is uninstalled. Open a command prompt and type “python”, if it throws an error, then your system has no python and your can proceed to freshly install python

4. Install python freshly

a.TF1.12 supports upto Python 3.6.6. Click here to download Windows x86-64 executable installer

b. While installing, select “Add Python 3.6 to PATH” and then click “Install Now”.

c. After successful installation of python, the installation window provides an option for disabling path length limit which is one of the root-cause of Tensorflow build/Installation issues in Windows 10 environment. Click “Disable path length limit” and follow the instructions to complete the installation.

d. Verify whether python installed correctly. Open a command prompt and type “python”. It should show the version of Python.

5. Install Visual Studio

Visual Studio 2017 Community

a. Click the "Visual Studio Link" above.Download Visual Studio 2017 Community.

b. Under “Visual Studio IDE” on the left, select “community 2017” and download it

c. During installation, Select “Desktop development with C++” and install

6. CUDA 9.0 toolkit

https://developer.nvidia.com/cuda-90-download-archive?target_os=Windows&target_arch=x86_64&target_version=10&target_type=exelocal

a. Click "Link to CUDA 9.0 toolkit" above, download “Base Installer”

b. Install CUDA 9.0

7. Install cuDNN

https://developer.nvidia.com/cudnn

a. Click "Link to Install cuDNN" and select “I Agree To the Terms of the cuDNN Software License Agreement”

b. Register for login, check your email to verify email address

c. Click “cuDNN Download” and fill a short survey to reach “cuDNN Download” page

d. Select “ I Agree To the Terms of the cuDNN Software License Agreement”

e. Select “Download cuDNN v7.5.0 (Feb 21, 2019), for CUDA 9.0"

f. In the dropdown, click “cuDNN Library for Windows 10” and download

g. Go to the folder where the file was downloaded, extract the files

h. Add three folders (bin, include, lib) inside the extracted file to environment

i. Type “environment” in windows 10 search bar and locate the “Environment Variables” and click “Path” in “User variable” section and click “Edit” and then select “New” and add those three paths to three “cuda” folders

j. Close the “Environmental Variables” window.

8. Install tensorflow-gpu

a. Open a command prompt and type “pip install --upgrade tensorflow-gpu”

b. It will install tensorflow-gpu

9. Check whether it was correctly installed or not

a. Type “python” at the command prompt

b. Type “import tensorflow as tf

c. hello=tf.constant(‘Hello World!’)

d. sess=tf.Session()

e. print(sess.run(hello)) -->Hello World!

10. Test whether tensorflow is using GPU

a. from tensorflow.python.client import device_lib print(device_lib.list_local_devices())

b. print(device_lib.list_local_devices())

Python 3.7 works for me, I uninstalled python 3.8.1 and reinstalled 3.7.6. After that, I executed:

pip3 install --user --upgrade tensorflow

and it works

if you tried the solutions above and didin't solve the problem, it can be because of version inconsistency.

I installed python 3.9 and i couldn't install tensorflow with pip.

And then I uninstalled 3.9, then installed 3.8.7 and success... the max version that tensorflow is supported by is 3.8.x (in 2021) so, check your python version is compatible or not with current tensorflow.

The URL to install TensorFlow in Windows, below is the URL. It worked fine for me.

python -m pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl

If you run into this issue recently (say, after Python 3.7 release in 2018), most likely this is caused by the lack of Python 3.7 support (yet) from the tensorflow side. Try using Python 3.6 instead if you don't mind. There are some tricks you can find from https://github.com/tensorflow/tensorflow/issues/20444, but use them at your own risk. I used the one harpone suggested - first downloaded the tensorflow wheel for Python 3.6 and then renamed it manually...

cp tensorflow-1.11.0-cp36-cp36m-linux_x86_64.whl tensorflow-1.11.0-cp37-cp37m-linux_x86_64.whl
pip install tensorflow-1.11.0-cp37-cp37m-linux_x86_64.whl

The good news is that there is a pull request for 3.7 support already. Hope it will be released soon.

I was facing the same issue. I tried the following and it worked. installing for Mac OS X, anaconda python 2.7

pip uninstall tensorflow export TF_BINARY_URL=<get the correct url from http://tflearn.org/installation/> pip install --upgrade $TF_BINARY_URL

Installed tensorflow-1.0.0