This project was a part of the Udacity Deep Learning Nanodegree. Here I train an LSTM model on a small part of the Simpsons dataset. I train the model on small sequence length (of sequence length of 2), where given a word the odel tries to predict he next word in the sequence. The final data generated by the model generates the punctuations, expressions or actions used while saying the dialogues quite decently.
Some packages required by the project is
- numpy
- Collections
- tensorflow=1.1.0