Predicting age and gender with AutoKeras
In this recipe, we'll study a practical application of AutoML that can be used as a template to create prototypes, MVPs, or just to tackle real-world applications with the help of AutoML.
More concretely, we'll create an age and gender classification program with a twist: the architecture of both the gender and age classifiers will be the responsibility of AutoKeras. We'll be in charge of getting and shaping the data, as well as creating the framework to test the solution on our own images.
I hope you're ready because we are about to begin!
Getting ready
We need a couple of external libraries, such as OpenCV, scikit-learn, and imutils. All these dependencies can be installed at once, as follows:
$> pip install opencv-contrib-python scikit-learn imutils
On the data side, we'll use the Adience dataset, which contains 26,580 images of 2,284 subjects, along with their gender and age. To download the...
