上QQ阅读APP看书，第一时间看更新
Technical requirements
- Python 3.6x 64-bit from https://www.python.org/
- NumPy for Python 3.6x
- TensorFlow from https://deepmind.com/ with TensorBoard
Programs: GitHub Chapter05.
The Python programs delivered with this chapter constitute a sandbox for you to play around with to become familiar with TensorFlow and TensorBoard. Feel free to run them, modify the TensorFlow data flow graphs, and see what happens in TensorBoard.
- FNN_XOR_Tensorflow.py: This program shows how to train an FNN to solve the XOR problem in a few lines.
- FNN_XOR_Tensorflow_graph.py: This program adds TensorBoard metadata to display the data flow graph of the program.
- FNN_XOR_Tensorflow_graph_only.py: This program contains a data flow graph of the program just to show that mastering the graph means mastering the program.
- Tensorboard.reader.py: This program runs with the preceding programs. It finds the /log file containing the data to display in TensorBoard and runs the TensorBoard local service.
- FNN_XOR_Tensorflow_tensorboard_MODEL1.py: This program shows how to tweak a FNN_XOR_Tensorflow_graph_only.py to add some labels to prepare a corporate presentation.
Check out the following video to see the code in action: