Install TensorFlow using 'tensorflow/tensorflow' Docker image

Start the Docker virtual machine

Create a virtual machine if not already

bash$ docker-machine create default

Start the virtual machine default:

bash$ docker-machine start
Starting "default"...
(default) Check network to re-create if needed...
(default) Waiting for an IP...
Machine "default" was started.
Waiting for SSH to be available...
Detecting the provisioner...
Started machines may have new IP addresses. You may need to re-run the `docker-machine env` command.
>>> elapsed time 35s
bash$ docker-machine env
export DOCKER_HOST="tcp://"
export DOCKER_CERT_PATH="/Users/meng/.docker/machine/machines/default"
export DOCKER_MACHINE_NAME="default"
# Run this command to configure your shell:
# eval $(docker-machine env)
bash$ eval "$(docker-machine env default)"

Start 'tensorflow/tensorflow' Docker image

$ export MY_WORKSPACE_DIR='/Users/meng/workspace'
$ docker run -it \
--net=host \
--publish 6006:6006 \
--volume ${MY_WORKSPACE_DIR}/tensorflow_test:/tensorflow_test \
--workdir /tensorflow_test \
tensorflow/tensorflow:1.1.0 bash
root@30d79c2e5fc3:/tensorflow_test# pwd

Start TensorBoard

root@30d79c2e5fc3:/tensorflow_test# tensorboard --logdir training_summaries &
[1] 12
root@30d79c2e5fc3:/tensorflow_test# Starting TensorBoard 47 at
(Press CTRL+C to quit)


Open TensorBoard at

in a Web browser, where the IP is the same as DOCKER_HOST in the output of docker-machine env.

Download training data

# curl -O
# tar xzf flower_photos.tgz
# find flower_photos -type f | wc -l
# cp -R flower_photos flower_photos_subset
# rm flower_photos_subset/*/[2-9]*
# find flower_photos_subset/ -type f | wc -l

Download training script

# curl -O


# python \
      --bottleneck_dir=bottlenecks \
      --how_many_training_steps=500 \
      --model_dir=inception \
      --summaries_dir=training_summaries/basic \
      --output_graph=retrained_graph.pb \
      --output_labels=retrained_labels.txt \

Python script to classify new images of flowers

Create a new file

import os, sys

import tensorflow as tf

os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'

# change this as you see fit
image_path = sys.argv[1]

# Read in the image_data
image_data = tf.gfile.FastGFile(image_path, 'rb').read()

# Loads label file, strips off carriage return
label_lines = [line.rstrip() for line 
                   in tf.gfile.GFile("retrained_labels.txt")]

# Unpersists graph from file
with tf.gfile.FastGFile("retrained_graph.pb", 'rb') as f:
    graph_def = tf.GraphDef()
    tf.import_graph_def(graph_def, name='')

with tf.Session() as sess:
    # Feed the image_data as input to the graph and get first prediction
    softmax_tensor = sess.graph.get_tensor_by_name('final_result:0')

    predictions =, \
             {'DecodeJpeg/contents:0': image_data})

    # Sort to show labels of first prediction in order of confidence
    top_k = predictions[0].argsort()[-len(predictions[0]):][::-1]

    for node_id in top_k:
        human_string = label_lines[node_id]
        score = predictions[0][node_id]
        print('%s (score = %.5f)' % (human_string, score))

Classify a new image:

# python flower_photos/daisy/99306615_739eb94b9e_m.jpg
daisy (score = 0.47996)
dandelion (score = 0.24667)
sunflowers (score = 0.24508)
tulips (score = 0.02018)
roses (score = 0.00811)

Exit the Docker image and stop the Docker virtual machine

Ctrl+D to exit Docker image and then run

$ docker-machine stop    


blog comments powered by Disqus