LiveKit Community

turn_detector is not perfect in french. any way to improve it ?

Thanks for reaching out, Antoine. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.



To improve the performance of the `turn_detector` for French, you can use the multilingual turn detection model. This model is based on Qwen2.5-0.5B and has excellent performance for French, with a 98.8% true positive rate and a 97.3% true negative rate.
To implement the multilingual model, you can use the following code snippet:
```from livekit.plugins.turn_detector.multilingual import MultilingualModel

session = AgentSession(
    ...
    turn_detection=MultilingualModel(),
)```
Note that the multilingual model requires approximately 400MB of RAM and completes inferences in about 25ms. This is slightly larger and slower compared to the English-only model, but it provides enhanced support for French and other languages.
 Sources: <https://docs.livekit.io/agents/build/turn-detection/|Turn detection and interruptions | LiveKit Docs> | <https://livekit.github.io/dse-workbench/output/agents-main.html|Tree View: | dse-workbench>

I'm sorry, but I don't have the information you need at the moment. If you would like a member of the LiveKit team to respond, please ask your question in one of the other channels.
