类class
agentscope.realtimeParameters
class Parameters(RealtimeModelBase.Parameters)源码Source
agentscope/realtime/_dashscope/_audio_model.py:24–72参数，处理低延迟的实时语音与多模态交互。位于 agentscope.realtime 包，继承自 RealtimeModelBase.Parameters。
Tuneables surfaced to the UI. Declared in agentscope.realtime. Extends RealtimeModelBase.Parameters.
属性摘要Attribute Summary
| 属性Attribute | 类型Type | 说明Description | 值Value |
|---|---|---|---|
voice | str | 语音字段。 | Field(default='longanqian', title='Voice', description='The voice used for spoken model responses.') |
turn_detection | Literal['server_vad', 'smart_turn', 'none'] | `turn_detection` 字段。 | Field(default='server_vad', title='Turn Detection', description='How the model detects when the user has finished speaking.') |
vad_threshold | float | 语音活动检测阈值字段。 | Field(default=0.5, ge=-1.0, le=1.0, description='Speech detection sensitivity. Lower values detect quieter speech.') |
vad_silence_duration_ms | int | `vad_silence_duration_ms` 字段。 | Field(default=800, ge=200, le=6000, description="Silence required before the model considers the user's turn complete.") |
voiceprint_audio_urls | list[str] | `voiceprint_audio_urls` 字段。 | Field(default_factory=list, title='Voiceprint Audio URLs', description='Reference audio URLs used by smart_turn to identify the intended speaker.') |
max_history_turns | int | 最大历史轮次字段。 | Field(default=20, ge=1, le=50, description='Maximum number of conversation turns retained by the model.') |
属性详情Attribute Detail
voice
str = Field(default='longanqian', title='Voice', description='The voice used for spoken model responses.')语音字段。声明为 str。
The voice used for spoken model responses.
源码Source
agentscope/realtime/_dashscope/_audio_model.py:27turn_detection
Literal['server_vad', 'smart_turn', 'none'] = Field(default='server_vad', title='Turn Detection', description='How the model detects when the user has finished speaking.')turn_detection 字段。声明为 Literal['server_vad', 'smart_turn', 'none']。
How the model detects when the user has finished speaking.
源码Source
agentscope/realtime/_dashscope/_audio_model.py:32vad_threshold
float = Field(default=0.5, ge=-1.0, le=1.0, description='Speech detection sensitivity. Lower values detect quieter speech.')语音活动检测阈值字段。声明为 float。
Speech detection sensitivity.
源码Source
agentscope/realtime/_dashscope/_audio_model.py:39vad_silence_duration_ms
int = Field(default=800, ge=200, le=6000, description="Silence required before the model considers the user's turn complete.")vad_silence_duration_ms 字段。声明为 int。
Silence required before the model considers the user's turn complete.
源码Source
agentscope/realtime/_dashscope/_audio_model.py:48voiceprint_audio_urls
list[str] = Field(default_factory=list, title='Voiceprint Audio URLs', description='Reference audio URLs used by smart_turn to identify the intended speaker.')voiceprint_audio_urls 字段。声明为 list[str]。
Reference audio URLs used by smart_turn to identify the intended speaker.
源码Source
agentscope/realtime/_dashscope/_audio_model.py:57max_history_turns
int = Field(default=20, ge=1, le=50, description='Maximum number of conversation turns retained by the model.')最大历史轮次字段。声明为 int。
Maximum number of conversation turns retained by the model.
源码Source
agentscope/realtime/_dashscope/_audio_model.py:65