Datasets:

---

MERaLiON

/

sea_audiobench_datasets_ST

Like 0

Follow

MERaLiON 131

Adds a dataset card for the ST (speech translation) task of SEA-SpeechBench.

Contents:

• YAML metadata: license, 9 source languages, task categories, and source dataset
• Loading example for datasets>=4.0
• Subset table: 9 configs, 7,189 examples, source corpus and audio duration per config
• Data field descriptions, including language_target
• Source data table with upstream license
• BibTeX for the benchmark and source corpus