Inspiration
We originally started with the idea of taking an author's works, and translating any arbitrary message into an author's tone and voice.
What it does
What it does now is translate between normal English, King James English, and Lolcat speak using OpenNMT
How we built it
We used OpenNMT, feeding each corpus of the NIV Bible, the King James Bible, and the Lolcat Bible to train the network.
Challenges we ran into
The original idea of using an author's works wasn't compatible with the form of translation that OpenNMT is based on. We needed a sentence-to-sentence comparison between the "languages" for the network to be properly trained, hence why we chose the Bible as a corpus.
Accomplishments that we're proud of
Parsing the sheer amount of data, and cleaning it such that it was usable by OpenNMT.
What we learned
How to more efficiently parse large text files. How to train OpenNMT.
What's next for King Lolcat Translator
Translating books in Project Gutenberg into Lolcat speak, and create a library of translated books.
Built With
- opennmt
- node.js
Log in or sign up for Devpost to join the conversation.