My projects from the Udacity Deep Learning Nanodegree.
Simple feed forward neural network that optimizes the number of bikes a bike rental service business should have in stock based on several data points ranging from weather to demand.
Convolutional neural network that determines the breed of a dog when given an image and can also identify humans. The model was trained on a set of images of dogs and humans.
Recurrent neural network with LSTM that generates new TV scripts after learning about them with a script dataset.
Generative Adversarial Network that generates celebrity images after being trained on the celeb image dataset.
RNN implemented on pytorch and deployed on AWS sagemaker (API accessible via webpage). Also trained an xgboost model. The models are trained on the IMDB dataset. App predicts sentiment of a movie review.