arioso.platforms.riffusion.adapter#

Riffusion adapter using the riffusion library or diffusers fallback.

Classes

Adapter(config)

Riffusion adapter with dual backend support.
class arioso.platforms.riffusion.adapter.Adapter(config)[source]#

Bases: object

Riffusion adapter with dual backend support.

Tries the riffusion library first for full spectrogram-to-audio support. Falls back to the diffusers StableDiffusion pipeline which generates a spectrogram image and converts it to audio via inverse STFT.

generate(prompt, *, negative_prompt='', seed=None, guidance=7.0, num_steps=50, audio_input_strength=0.75, **kwargs)[source]#

Generate audio from a text prompt via spectrogram synthesis.

Parameters:

  • prompt (str) – Text description of desired music.

  • negative_prompt (str) – Elements to avoid in generation.

  • seed (int) – Random seed for reproducibility.

  • guidance (float) – Classifier-free guidance scale.

  • num_steps (int) – Number of diffusion inference steps.

  • audio_input_strength (float) – Denoising strength (0-1).

Return type:

Song

Returns:

A Song with audio_array populated.