Getting started with TensorFlow on Windows

I have wanted to try using TensorFlow for a while since it is very popular, and I think it might be able to solve a problem I’m working on this summer. I have a Windows 7 laptop, so I’ll be detailing the installation process for TensorFlow on a Windows computer.

This is mostly for my own reference purposes, you will learn nothing by reading this. I’m just following the details provided by TensorFlow here https://www.tensorflow.org/install/install_windows.

I have Anaconda installed with Python 2.7. TensorFlow needs Python 3.5, so I opened an Anaconda Prompt and entered conda create -n tensorflow python=3.5. This created a new environment with Python 3.5. To switch to this environment, I entered activate tensorflow. Then I installed the CPU-only version of TF using the command

(tensorflow)C:> pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/windows/cpu/tensorflow-1.2.1-cp35-cp35m-win_amd64.whl

In the same prompt I entered python to start running Python. Then the following lines loaded TF.

>>> import tensorflow as tf
>>> hello = tf.constant('Hello, TensorFlow!')
>>> sess = tf.Session()
>>> print(sess.run(hello))

I got the expected result, so it seems that TensorFlow was set up properly.

To make sure it is working I’ll run the linear model example found on https://www.tensorflow.org/get_started/get_started. I copied the following code into my prompt

import numpy as np
import tensorflow as tf

# Model parameters
W = tf.Variable([.3], dtype=tf.float32)
b = tf.Variable([-.3], dtype=tf.float32)
# Model input and output
x = tf.placeholder(tf.float32)
linear_model = W * x + b
y = tf.placeholder(tf.float32)
# loss
loss = tf.reduce_sum(tf.square(linear_model - y)) # sum of the squares
# optimizer
optimizer = tf.train.GradientDescentOptimizer(0.01)
train = optimizer.minimize(loss)
# training data
x_train = [1,2,3,4]
y_train = [0,-1,-2,-3]
# training loop
init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init) # reset values to wrong
for i in range(1000):
  sess.run(train, {x:x_train, y:y_train})

# evaluate training accuracy
curr_W, curr_b, curr_loss = sess.run([W, b, loss], {x:x_train, y:y_train})
print("W: %s b: %s loss: %s"%(curr_W, curr_b, curr_loss))

The output is

W: [-0.9999969] b: [ 0.99999082] loss: 5.69997e-11

Exactly what the website said I should get.

It looks like it everything is working fine. Most surprisingly it all worked on the first try, which is a rare occurance for getting ML software to work on Windows.