Is Tensorflow compatible with a Windows workflow?
Updated 11/28/2016: Today we released the first release candidate of TensorFlow 0.12, which includes support for Windows. You can install the Python bindings using the following command in a Python shell:
C:\> pip install tensorflow
...or, if you want GPU support:
C:\> pip install tensorflow-gpu
You can also build TensorFlow yourself using Microsoft Visual C++ and NVCC (for the CUDA parts). The easiest way to build on Windows is currently to use the CMake build, and we will soon provide support for Bazel on Windows.
Previous answer: We haven't tried to build TensorFlow on Windows so far: the only supported platforms are Linux (Ubuntu) and Mac OS X, and we've only built binaries for those platforms.
For now, on Windows, the easiest way to get started with TensorFlow would be to use Docker: http://tensorflow.org/get_started/os_setup.md#docker-based_installation
It should become easier to add Windows support when Bazel (the build system we are using) adds support for building on Windows, which is on the roadmap for Bazel 0.3. You can see the full Bazel roadmap here.
In the meantime, you can follow issue 17 on the TensorFlow GitHub page.
Install tensorflow on Windows with anaconda
UPDATE: Since TensorFlow 0.12, we have published packages for Windows. You can install the CPU-only version with the following command:
C:\> pip install tensorflow
…and the GPU-accelerated version with:
C:\> pip install tensorflow-gpu
Note that you will need the 64-bit version of Python 3.5 installed for the above commands to work.
TensorFlow is not currently supported on Windows, and none of the official binary packages work on Windows. We are currently working on adding support for Windows, but this effort is in the early stages.
See the answers to this question for suggestions on how to run TensorFlow using Docker or Bash for Windows.
Using created Tensorflow CNN models in Windows?
The only way to run TensorFlow on Windows now is using a VM/Docker, steps outlined here.
However, if you are just interested in running inferences (without training/updating your model), you could take a look at tfdeploy which exports your TensorFlow code to Numpy code, so that you only depend on Numpy.
TensorFlow in Visual Studio 2015, using Canopy as the Python environment
The current version of TensorFlow does not support running on Windows. In particular, the PIP packages available here are for Linux and Mac OS X only: they include a native compiled library called _pywrap_tensorflow
that has not been compiled for Windows.
As suggested in another answer, you can run TensorFlow on Windows using a Docker container. There is an open issue on GitHub concerning better Windows support, but this will depend on Bazel adding support for Windows build (currently targeted for later this year).
How can one validate a Tensor Flow installation on Windows
TensorFlow versions 1.1.0 and higher have been compiled for Python 3.6 (as well as 3.5 in most cases).
You can check the current installation of TensorFlow using command:
python -c "import tensorflow as tf; print(tf.__version__)"
Related Topics
Using a Python Subprocess Call to Invoke a Python Script
Weighted Choice Short and Simple
Comments Not Working in Jinja2
Why Are Default Arguments Evaluated at Definition Time
Get List of All Routes Defined in the Flask App
Pandas: Convert Dtype 'Object' to Int
Cannot Display an Image in Tkinter
Simple Animation Using Tkinter
Smtplib Sends Blank Message If the Message Contain Certain Characters
Why Does Indexing Numpy Arrays with Brackets and Commas Differ in Behavior
Where to Put Django Startup Code
Python: How to Run Eval() in the Local Scope of a Function
How to Leave/Exit/Deactivate a Python Virtualenv
Differencebetween a Pandas Series and a Single-Column Dataframe
Panda's Dataframe - Renaming Multiple Identically Named Columns
Minimum Euclidean Distance Between Points in Two Different Numpy Arrays, Not Within