The goal of the singularizer is to take one skittle from a group of skittles. The reason we need this done is so that we can scan the color and move it solely to the correct bin. If we didn't have the singularizer all the skittles would fall into 1 bin regardless of color. Our first attempt at making a singularizer didn't work very well. we only used 1 piece of 1/8 thick cardboard so the weight of a skittle made it bend allowing multiple skittles to fall through. After that we changed the shape and made it 3 cardboard pieces thick for extra durability. Now it works successfully 100% of the time and does its job great.

The goal is to make a model that can learn based on images of people's faces and use that information to recognize their faces with an accuracy of 80% or higher

There are two data sets being used to train the model, train_images and train_labels. These datasets have loads of pictures and labels and when ran it trains the computer based off of what is in the datasets.

The model takes a photo of your face and attempts to recognize it using the images and labels it was trained on. Once it thinks it have predicted who you are it says your name through a speaker.

My model has an accuracy of 89% using 80 epochs. This exceeds my goal of having at least 80% accuracy.