A work applying recent NLP techniques involving word embeddings and sequential models to leverage previously unused data in classifying Library of Congress records, improving performance despite the small available data.
An RNN that generates text which closely resembles the writing style of Shakespeare. Implemented using PyTorch.
A neural network that can turn your grayscale images of pets into colored ones. Built using the [Fast.ai](https://www.fast.ai/) library.
A minimal C++ implementation of the famous SuperMario game, using [SDL2](https://www.libsdl.org/). This was a course project for the Advanced Programming course at the University of Tehran, and I was the TA designing the project. The project would be shown to students as an example of good design.
A neural network for the classification of letters A-J from the [notMNIST](http://yaroslavvb.blogspot.com/2011/09/notmnist-dataset.html) dataset, built from scratch using Python and without the use of any external libraries.