(inspired by TensorFlow Examples)
This example illustrates the processes to train handwriting recognition model using a hand-coded multiple layers neural network on numpy.
In our examples, we are using input_data.py given by the tutorial of TensorFlow to load that dataset. It is quite useful for managing our data, and handle:
- Dataset downloading
- Loading the entire dataset into numpy array
This example also shows a fundamental application of neural network. There are several techniques had not been included such as hyperparameter tunning, model optimization etc. However, readers with basic knowledge about Deep Learning can easily gain an intuition of how multi-layer neural network works through this example.