Uses a character-level language model, to learn from a million examples and thereafter generate new names on its own.
A few of them after a few iterations are:
ethreen
blenn
ginnie
lyvina
hilford
roza
gerraro
The model uses an RNN with a basic RNN unit, implemented from scratch using python. The test dataset has been taken from a kaggle competition.
The file nameGenerator_MBGD.py uses a mini-batch GD for training, while the nameGenerator_SGD.py uses stochastic GD.