LLMUserContextAggregator: ignore short uterrances while bot speaking
This commit is contained in:
@@ -25,6 +25,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
|
||||
|
||||
### Changed
|
||||
|
||||
- Short utterances not detected by VAD while the bot is speaking are now
|
||||
ignored. This reduces the amount of bot interruptions significantly providing
|
||||
a more natural conversation experience.
|
||||
|
||||
- Updated `GladiaSTTService` to output a `TranslationFrame` when specifying a
|
||||
`translation` and `translation_config`.
|
||||
|
||||
|
||||
Reference in New Issue
Block a user