The open-source language model was trained using Wikipedia text and transcripts of over 7,000 speeches.
10 Jun 2019 10:25 AM IST