Skip to content
Fan Zeng (Frank)
Back to resources

Speech-MASSIVE

Speakers record short voice-assistant requests with matching prompt text, intent and slot labels.

Visit official source (opens in a new tab)

Research use

Compare command recognition, slot preservation and few-shot adaptation across languages.

Limitations

Scripted requests cover a narrower range of speaking styles than conversation. The full training split is offered for French and German.

License and access

Dataset license
CC BY-NC-SA 4.0 (opens in a new tab)
Access status
Direct access

Select a language or the combined configuration. The released Parquet files contain audio and text.

Additional sources