Struct FTryllVoiceInputConfig¶
ClassList > FTryllVoiceInputConfig
#include <TryllVoiceInput.h>
Public Attributes¶
| Type | Name |
|---|---|
| float | HotwordsScore = 1.5f |
| FString | HotwordsStoragePath |
| FTryllAudioFormat | InputFormat |
| FString | ModelName |
| uint32 | VadMinSilenceMs = 1000 |
| uint32 | VadSpeechPadMs = 250 |
| float | VadThreshold = 0.5f |
Detailed Description¶
Parameters for creating an FTryllVoiceInput handle.
Public Attributes Documentation¶
variable HotwordsScore¶
How strongly matched phrases are weighted against one another during beam search (default 1.5). Measured on Parakeet TDT 0.6B v3: 1.8 recovered 5 of 8 target terms, 2.5 recovered 6 of 8. A lexicon line may override this for itself by ending with :<number>.
variable HotwordsStoragePath¶
Relative path (resolved against the session storage folder) to a StringStorage (kind=List) config whose phrases the STT decoder biases toward. Loaded on demand. Empty (default) disables hotword biasing.
variable InputFormat¶
variable ModelName¶
variable VadMinSilenceMs¶
variable VadSpeechPadMs¶
variable VadThreshold¶
The documentation for this class was generated from the following file C:/_tryll/_monorepo3/tryll/clients/unreal/Source/TryllClient/Public/TryllVoiceInput.h