Soft Window Transformers for Handwriting Synthesis

Adaptation of the soft window mechanism to transformers and applying it to handwriting synthesis

Double DQN agent for Atari pong

Implementation of double DQN agent for the atari pong game, discussing libraries such as pfrl and torchrl

Handwriting generation with RNNs

Combining RNNs with a mixture of gaussians to teach an algorithm how to handwrite.

Neural Style Transfer

Streamlit application for Neural Style Transfer using Keras and Tensorflow.

Pendulum RL

Reinforcement Learning for pendulum equilibrium using numpy and value function iteration.