makemore-CityNames

makemore-CityNames takes in a single .csv file filled with 20000+ US City Names from this Kaggle dataset and uses machine learning models to predict new ones.

This repository was inspired by and follows along with Andrej Karpathy's Neural Networks: Zero to Hero youtube playlist.

Models implemented:

• Bigram and Trigram - Character-Level Language Model
• Multi-Layer Perceptron (MLP) - Character-Level Language model
• Recurrent Neural Network (RNN) - Character-Level Language model (kind of like WaveNet)

Other implementations:

• Manual Backpropogation