Fact · Speech recognition
10000hours
GigaSpeech is an English speech recognition corpus with 10,000 hours of high quality labeled audio, covering several domains.
Checked against its source on October 17, 2026
In the source's own words
“an evolving, multi-domain English speech recognition corpus with 10,000 hours of high quality labeled audio”
GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audio, Guoguo Chen et al., arXiv, 2021, Abstract.
Open the original- What it measures
- The corpus was assembled from audiobooks, podcasts and YouTube videos, covering read and spontaneous speech. Its training data is split into subsets from 10 hours to the full 10,000.
- What it does not show
- The figure is the size of the transcribed training data, not the number of hours in any single test set. The audio's terms of use are a separate matter, covered by the dataset card.