Taiwan Tongues ASR CE — Hakka
The Hakka release pairs speech files with reference sentences and duration metadata for each recording.
Visit official source (opens in a new tab)Research use
Evaluate low-resource recognition and use the supplied durations to form subsets with a fixed audio budget.
Limitations
The publisher does not give a complete breakdown by dialect or recording condition.
License and access
- Dataset license
- TRAIL-D dataset terms (opens in a new tab)
- Access status
- Direct access
Download the train or test WebDataset archives. The corresponding TSV maps file identifiers to text.