This is the third project of the Udacity Deep Learning Nanodegree. In this project, I built a TV scripts generator using RNNs. Trained on part of the Seinfeld dataset of scripts from 9 seasons, the RNN can generate a new, "fake" TV script.
Model training in this project uses the Seinfeld Chronicles dataset, which is publicly available via Kaggle. This dataset is ~3.4MB.
The model uses recurrent neural networks (RNN) and consists of an embedding layer (with 256 embedding dimensions), a 2-layer LSTM (with 512 hidden dimensions), and, finally, a fully-connected layer. With 15 epochs of training, the RNN was able to achieve ~2.7-2.8 in training loss. Lastly, the RNN can generate "fake" TV scripts of any specified length when a prime word is given.
-
Clone the repository and navigate to the downloaded folder.
git clone https://github.com/chloeh13q/DLND-TV-Script-Generation cd DLND-TV-Script-Generation
-
Make sure you have already installed the necessary Python packages.
-
Open a terminal window and navigate to the project folder. Open the notebook and follow the instructions.
jupyter notebook dlnd_tv_script_generation.ipynb