The final source code with TensorFlow and TensorBoard

After running the programs, to make sure nothing is missing, the TensorFlow and TensorBoard models can be merged and simplified as shown in FNN_XOR_Tensorflow_graph.py.

The summaries have been simplified, the cost function uses the optimized least squares method, and XOR linear separability has been proven with this model.