Description: 10,000 hours of English audio from audiobooks, podcasts, and YouTube. Designed as a large-scale ASR training corpus. - License: Apache 2.0. - Chapters: 24.