For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Python class

OutputAudioContent

OutputAudioContent

class max.pipelines.request.OutputAudioContent(*, type='output_audio', audio_url=None, audio_data=None, format=None, sample_rate=None, num_samples=None, samples=None)

source

Bases: BaseModel

Audio content generated by the model in output messages.

Parameters:

• type (Literal['output_audio'])
• audio_url (str | None)
• audio_data (str | None)
• format (str | None)
• sample_rate (int | None)
• num_samples (int | None)
• samples (Annotated[ndarray[tuple[Any, ...], dtype[float32]] | None, SkipJsonSchema()])

audio_data

audio_data: str | None

source

audio_url

audio_url: str | None

source

format

format: str | None

source

from_numpy_samples()

classmethod from_numpy_samples(samples, *, sample_rate, format=None)

source

Creates internal audio content from a raw waveform.

Parameters:

• samples (ndarray[tuple[Any, ...], dtype[float32]]) – A float32 array with shape [channels, samples] holding values in [-1, 1].
• sample_rate (int) – Sample rate of the waveform, in hertz.
• format (str | None) – Optional container metadata, typically "wav".

Returns:

An OutputAudioContent carrying the raw waveform, for the route to encode into a container.

Raises:

ValueError – If the dtype, rank, or sample rate is invalid.

Return type:

OutputAudioContent

model_config

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True, 'frozen': True}

source

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

num_samples

num_samples: int | None

source

sample_rate

sample_rate: int | None

source

samples

samples: SkipJsonSchema[npt.NDArray[np.float32] | None]

source

type

type: Literal['output_audio']

source