open-speech-corpora
open-speech-corpora copied to clipboard
kreyòl ayisyen :)
http://www.speech.cs.cmu.edu/haitian/speech
haitian creole from CMU
from linkedin "In 2010, we released both the text data & speech models. Everyone was more focused on the text data to create MT systems, yet we had also created both speech-to-text and text-to-speech in a vi-directional speech-to-speech MT system. All of the Haitian data was carefully collected and curated over 2+ years. The project funding ended in 1999, and the system has been on a shelf ever since. Thus, we made the data and models publicly available so others could benefit from it. The free licensing terms are indicated on directory up at that link."
What would it take to get this going?