Datasets:

---

MERaLiON

/

sea_audiobench_datasets_ASR

Like 1

Follow

MERaLiON 131

Adds a dataset card for the ASR (automatic speech recognition) task of SEA-SpeechBench.

Contents:

• YAML metadata: license, 11 languages, task categories, source datasets, and gated-access notice for the non-commercial source material
• Loading example for datasets>=4.0
• Subset table: 33 configs, 26,863 examples, source corpus and audio duration per config
• Data field descriptions
• Source data table with upstream licenses for all 15 source corpora
• BibTeX for the benchmark and all source corpora