类class agentscope.asr

OpenAIASRModel

class OpenAIASRModel(ASRModelBase)

源码Source agentscope/asr/_openai/_model.py:14–72

OpenAI ASR 模型，处理低延迟的实时语音与多模态交互。位于 agentscope.asr 包，继承自 ASRModelBase。核心成员包括 transcribe。

Batch speech recognition using OpenAI's Audio Transcriptions API. Declared in agentscope.asr. Extends ASRModelBase.

嵌套类摘要Nested Class Summary

类Class	类型Kind	说明Description
`Parameters`	类class	参数，处理低延迟的实时语音与多模态交互。 Options supported by OpenAI transcription models.

构造器Constructor	说明Description
`def __init__(credential:` `OpenAICredential, model: str = 'gpt-4o-mini-transcribe', parameters: 'OpenAIASRModel.Parameter…`	初始化 `OpenAIASRModel` 实例。 Creates a new instance.

属性Attribute	类型Type	说明Description	值Value
`type`	Literal['openai_asr']	类型字段。	`'openai_asr'`
`realtime`	bool	实时字段。	`False`

方法Method	说明Description
`async def transcribe(audio: bytes \| None = None, filename: str = 'audio.wav', **kwargs: Any) ->` `ASRResponse`async	执行 `transcribe` 方法。 Transcribe in-memory audio, using ``filename`` for its format.

def __init__(credential: OpenAICredential, model: str = 'gpt-4o-mini-transcribe', parameters: 'OpenAIASRModel.Parameters | None' = None) -> None

初始化 OpenAIASRModel 实例。

Creates a new instance.

参数Parameters

名称Name	类型Type	说明Description
`credential`	OpenAICredential	凭证对象（类型 `OpenAICredential`）。The credential.
`model`默认default `'gpt-4o-mini-transcribe'`	str	模型实例（类型 `str`）。The model.
`parameters`默认default `None`	'OpenAIASRModel.Parameters \| None'	参数（类型 `'OpenAIASRModel.Parameters \| None'`）。The parameters.

源码Source agentscope/asr/_openai/_model.py:35–49

Literal['openai_asr'] = 'openai_asr'

类型字段。声明为 Literal['openai_asr']。

The type field.

源码Source agentscope/asr/_openai/_model.py:29

bool = False

实时字段。声明为 bool。

The realtime field.

源码Source agentscope/asr/_openai/_model.py:32

async def transcribe(audio: bytes | None = None, filename: str = 'audio.wav', **kwargs: Any) -> ASRResponse

async

执行 transcribe 方法。

Transcribe in-memory audio, using filename for its format.

参数Parameters

名称Name	类型Type	说明Description
`audio`默认default `None`	bytes \| None	音频（类型 `bytes \| None`）。The audio.
`filename`默认default `'audio.wav'`	str	filename（类型 `str`）。The filename.
`kwargs`	Any	关键字参数（类型 `Any`）。The kwargs.

返回Returns ASRResponse

返回ASRResponse。Returns ASR response.

源码Source agentscope/asr/_openai/_model.py:51–72

class Parameters

参数，处理低延迟的实时语音与多模态交互。位于 agentscope.asr 包。

Options supported by OpenAI transcription models. Declared in agentscope.asr.

继承自Inherited from ASRModelBase