Notebooks, data, and models for various projects I've worked on.

Some datasets and models are too large for Github to handle and live on my personal server instead. If you encounter any missing data, that's what's happened!

• data contains some datasets I play with.
• models contains model artifacts I've created.
• notebooks is where the action is.

A learning notebook where I introduce Artificial Neural Networks.

Experiments with a few different document classification strategies.

I use VGG16 for the lower layers of my model and add a few additional layers to perform emotion classification. Accuracy is only around 45% across 7 classes (random would be 14%).

I implement this paper which uses the same dataset as Emotion Detection v1 and creates an XCEPTION model. Accuracy is 65%.

I teach you literally everything about linear regression.

Experimenting with filters in Keras.

I create a CNN that classifies clothing. I then apply this model to clothing I found on Google images.

I show you how to regularize overfit models.

I use math to measure how similar different languages sound to each other - the results are surprising.