Project Overview

This repository hosts code used in third project on Recurrent Neural Networks (RNN) project in the Deep Learning Nanodegree.

The aim is to generate Seinfeld TV scripts using RNNs using Seinfeld dataset of scripts from 9 seasons. The Neural Network built generates a new, "fake" TV script.

Getting started

Open the tv_script_generation.ipynb on a GPU enabled Python notebook environment, or Google colab (a link is directly on top in the notebook file). The notebook consists of further technical details. Run the notebook to generate your own unique script.

Further improvemets

• Add validation data to choose the best model
• Initialize model weights, especially the weights of the embedded layer to encourage model convergence
• Use topk sampling to generate new words.
• Generate music like DeepBach.
• Predict seizures in intracranial EEG recordings on Kaggle.