Designing and training a NN model
In this recipe, we will be leveraging the following NN architecture to recognize our words:
The model has two two-dimensional (2D) convolution layers, one dropout layer, and one fully connected layer, followed by a softmax activation.
The network's input is the MFCC feature extracted from the 1-s audio sample.
Getting ready
To get ready for this recipe, we just need to know how to design and train a NN in Edge Impulse.
Depending on the learning block chosen, Edge Impulse exploits different underlying ML frameworks for training. For a classification learning block, the framework uses TensorFlow with Keras. The model design can be performed in two ways: