Using gpt_2_simple to create a song generator, scraping the web for songs. I chose wu-tang songs to train on because I am a fan of their music. There are many other variations to do this with, basically anything in written format. This repo is probably breaching some kind of copyright, sorry about that. If it is an issue let me know.
... Shogun King,
The A.I. covers you from the worst
Ticalah, greg, Caligula, body of the beast
Shogun,
Nah, I shot you, I killed you, ni**a
You don't know me, I'm the one that shot you, ni**a
You know what I mean?
Yeah, I know, I like the way you dress
Yeah, I like the way you look
Yeah, I'm the one that shot you, ni**a
You know what I mean?
Yeah, I know, I like the way ...
... Wu-Tang Gods was spotted in minineum, like the RZA
Chill scroll dusty dusty, the Wu-Tang walls seem ready
Death to the art I be considerous harm, the air that's laced with hazardous loony
gloss laced by five underground compels it to attract to common allure for Benz platinum
Triagram slug find my shipment full of trife that's not safe rhyme ...
You will need the following python libs:
- gpt_2_simple
- bs4
Start a Jupyter Notebook and run gather-lyrics first and then train your model.
WTFPL