arioso.platforms.harmonai.adapter#

Harmonai (Dance Diffusion) adapter using diffusers.

Classes

Adapter(config)

Dance Diffusion adapter with lazy model loading.
class arioso.platforms.harmonai.adapter.Adapter(config)[source]#

Bases: object

Dance Diffusion adapter with lazy model loading.

Uses the diffusers library to run unconditional audio generation via the Dance Diffusion pipeline.

generate(prompt='', *, num_steps=100, seed=None, batch_size=1, **kwargs)[source]#

Generate audio unconditionally.

Parameters:

  • prompt (str) – Ignored (unconditional model). A warning is emitted if a non-empty prompt is provided.

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

  • seed (int) – Random seed for reproducibility.

  • batch_size (int) – Number of audio samples to generate.

Return type:

Song

Returns:

A Song with audio_array populated.