ros_tensorflow tutorial

This repo implements a Tensorflow 2 integration in ROS. Explanations are given in the following blog post.

Files

This repo contains two ROS packages:

• ros_tensorflow
• ros_tensorflow_msgs

Installing

1. Install the dependencies: pip3 install -r requirements.txt
2. Clone this repo to your catkin workspace
3. Build the workspace: catkin_make

Running

1. Run the node: rosrun ros_tensorflow node.py
2. Predictions are performed at regular time interval in a loop
3. Train the Tensorflow model (use tab completion):

rostopic pub /train/goal ros_tensorflow_msgs/TrainActionGoal "header:
  seq: 0
  stamp:
    secs: 0
    nsecs: 0
  frame_id: ''
goal_id:
  stamp:
    secs: 0
    nsecs: 0
  id: ''
goal:
  epochs: 10"

5. Abort training (use tab completion):

rostopic pub /train/cancel actionlib_msgs/GoalID "stamp:
  secs: 0
  nsecs: 0
id: ''"

6. Run predicitons from a service call (use tab completion):

rosservice call /predict "data: [0.,0.,0.,0.,0.,0.,0.,0.,0.,0.]"