Difficulty: beginner
Estimated Time: 10 minutes

The previous scenario walked you through the process of creating and training the neural network. The simple architecture is not sufficient for more complex data. Here we will build a neural network that has the hidden layer to accomplish the task of classifying the data that are not linearly separable.


You've completed TensorFlow Deep Network Training scenario.

You've used the capabilities of TensorFlow to create the hidden layer and train the network on the non linearly separated data.

TensorFlow Deep Network Training

Step 1 of 5


In the last scenario, we've built a simple neural network, with just one output layer. As we know this is not sufficient when working with the non linearly separated data.


The network will need the hidden layer, so the architecture will look as follows:

Neural network

You will be writing the code in the neural_network.py file. Specific tasks have been marked. As you can see, we have loaded the data for you and set up some variables for the further process.

Task 1

The first thing you need to do is to define the placeholders for the input data and labels. Because the size of the input data doesn't have to be defined now, you can use None.