tensorflow_ros_test

A package for testing if tensorflow_ros_cpp works on your machine. Report all bugs there.

If you succeed to build it, try running

rosrun tensorflow_ros_test tensorflow_ros_test_node

It tests running models/train.pb with a pre-built tensorflow graph.

You can create such a graph e.g. by:

import tensorflow as tf
import numpy as np

with tf.Session() as sess:
    a = tf.Variable(5.0, name='a')
    b = tf.Variable(6.0, name='b')
    c = tf.multiply(a, b, name="c")

    sess.run(tf.initialize_all_variables())

    print a.eval() # 5.0
    print b.eval() # 6.0
    print c.eval() # 30.0

    tf.train.write_graph(sess.graph_def, 'models/', 'train.pb', as_text=False)

Thanks to Jim Fleming for the sample code.

Troubleshooting

If you run into any kind of compilation/linking errors, be sure to read tensorflow_ros_cpp README as there are many information about which combination of Tensorflow installation and system work together.

Specifically, if you get linking errors containing cxx11 words, it means your system is too new to use the pip-installed Tensorflow easily. Consult [https://github.com/tradr-project/tensorflow_ros_cpp#c-abi-difference-problems]. One solution is to hide all Tensorflow code behind a C API as shown here.