SmoothConv
SmoothConv pairs naturally occurring Mandarin dialogue with expert annotations of turn events and conversational context.
Visit official source (opens in a new tab)Research use
Use the manually reviewed labels to assess interaction timing or benchmark models trained on DuplexConv.
Limitations
The data cover particular conversational domains. Annotation decisions and collection conditions affect comparability.
License and access
- Dataset license
- CC BY-NC 4.0 (opens in a new tab)
- Access status
- Direct access
The publisher’s Hugging Face repository provides WAV files without an account. Follow the release terms for research use.