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

AudioGenerationDetails

AudioGenerationDetails

class max.pipelines.request.AudioGenerationDetails(*, duration_seconds, sample_rate, channels, num_samples, steps)

source

Bases: BaseModel

Audio generation usage metadata.

The audio counterpart of ImageGenerationDetails: what a caller would need to attribute usage, taken from the audio that was actually produced rather than from what the request asked for, since a model may stop before the duration it was given.

Parameters:

• duration_seconds (float)
• sample_rate (int)
• channels (int)
• num_samples (int)
• steps (int)

channels

channels: int

source

duration_seconds

duration_seconds: float

source

from_waveform()

classmethod from_waveform(waveform, *, sample_rate, steps)

source

Build audio generation details from a generated waveform.

Parameters:

• waveform (ndarray[tuple[Any, ...], dtype[float32]]) – The generated samples, shaped (channels, samples).
• sample_rate (int) – Sample rate of the waveform, in hertz.
• steps (int) – The number of denoising steps used to generate.

Returns:

An AudioGenerationDetails describing the generated audio.

Raises:

ValueError – If the waveform is not two-dimensional, or the sample rate is not positive.

Return type:

AudioGenerationDetails

model_config

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

source

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

num_samples

num_samples: int

source

sample_rate

sample_rate: int

source

steps

steps: int

source