Following along and playing with Andrej Karpathy's makemore language modelling course (https://www.youtube.com/playlist?list=PLAqhIrjkxbuWI23v9cThsA9GvCAUhRvKZ)

Modularised the code rather than purely working with notebooks. See notebooks for a run through of the various models.

Data is from the US baby names dataset, which contains names, genders, and counts of babies born with that name in each year from 1880-2022. It can be downloaded here: https://www.ssa.gov/oact/babynames/limits.html (use the national data link)

To setup, run pip install . from the root directory.

Models included: bigram.ipynb

• Bigram count-based model
• Bigram neural network model embeddings.ipynb: