An LSTM implementation for a Rap Lyric Generator that spawns rap lyrics based on a Kaggle dataset with over 38,000 lines.
In order to run the Generator, the following must be done after getting into the Generator's working directory:
- Unzip the dataset through the two following commands:
sudo apt-get install unzip
unzip lyrics.csv.zip
- Train the LSTM network through the following command:
python train.py
- Generate new lyrics through the following command:
python run.py -s <seed>
The 'seed' parser argument can be replaced for anything, such as a word, phrase or line. It is simply the starting point of generating the text.
The features of the Rap Lyrics Generator include:
- Ability to create lyrics based on a seed
- Generates 1000 character lyric outputs
- Trains efficiently
- Utilizes rap lyrics from various artists, such as Kanye West and Kendrick Lamar, to learn
The Rap Lyrics Generator requires the following to be installed on the system:
- Python 3
- Tensorflow
- Keras
- Numpy
- Pandas