Difficulty: beginner
Estimated Time: 10 minutes

It's time to use the knowledge from the last scenarios to build the Deep Neural Network to classify the MNIST Dataset. You will create more complex architecture to accomplish the task.


You've completed MNIST Dataset Deep Learning scenario.

You've used the capabilities of TensorFlow to create the hidden layer and train the network on the MNIST Dataset.

Don’t stop now! The next scenario will only take about 10 minutes to complete.

MNIST Dataset Deep Learning

Step 1 of 6

Load Dataset

The goal of the current scenario, you're going to build a deeper neural network. This means you will add a hidden layer before the output one.

We've prepared the backbone of the code for you in the neural_network.py file. The tasks have been pointed out.

Task 1

The first task is to lead the MNIST dataset. You can use the input_data.read_data_sets function. Use a path of your choice and the one_hot encoding.

The images are of the 28 x 28 size. The hidden layer has 1024 neurons and there are 10 labels.

image_size = 28 labels_size = 10 hidden_size = 1024