Dance Diffusion
Dance Diffusion is by Zach Evans.
Dance Diffusion is the first in a suite of generative audio tools for producers and musicians released by Harmonai.
The original codebase of this implementation can be found at Harmonai-org.
Make sure to check out the Schedulers guide to learn how to explore the tradeoff between scheduler speed and quality, and see the reuse components across pipelines section to learn how to efficiently load the same components into multiple pipelines.
DanceDiffusionPipeline
class diffusers.DanceDiffusionPipeline
< source >( unet scheduler )
Parameters
- unet (UNet1DModel) —
A
UNet1DModelto denoise the encoded audio. - scheduler (SchedulerMixin) —
A scheduler to be used in combination with
unetto denoise the encoded audio latents. Can be one of IPNDMScheduler.
Pipeline for audio generation.
This model inherits from DiffusionPipeline. Check the superclass documentation for the generic methods implemented for all pipelines (downloading, saving, running on a particular device, etc.).
__call__
< source >( batch_size: int = 1 num_inference_steps: int = 100 generator: typing.Union[torch._C.Generator, typing.List[torch._C.Generator], NoneType] = None audio_length_in_s: typing.Optional[float] = None return_dict: bool = True ) → AudioPipelineOutput or tuple
Parameters
- batch_size (
int, optional, defaults to 1) — The number of audio samples to generate. - num_inference_steps (
int, optional, defaults to 50) — The number of denoising steps. More denoising steps usually lead to a higher-quality audio sample at the expense of slower inference. - generator (
torch.Generator, optional) — Atorch.Generatorto make generation deterministic. - audio_length_in_s (
float, optional, defaults toself.unet.config.sample_size/self.unet.config.sample_rate) — The length of the generated audio sample in seconds. - return_dict (
bool, optional, defaults toTrue) — Whether or not to return a AudioPipelineOutput instead of a plain tuple.
Returns
AudioPipelineOutput or tuple
If return_dict is True, AudioPipelineOutput is returned, otherwise a tuple is
returned where the first element is a list with the generated audio.
The call function to the pipeline for generation.
Example:
from diffusers import DiffusionPipeline
from scipy.io.wavfile import write
model_id = "harmonai/maestro-150k"
pipe = DiffusionPipeline.from_pretrained(model_id)
pipe = pipe.to("cuda")
audios = pipe(audio_length_in_s=4.0).audios
# To save locally
for i, audio in enumerate(audios):
write(f"maestro_test_{i}.wav", pipe.unet.sample_rate, audio.transpose())
# To dislay in google colab
import IPython.display as ipd
for audio in audios:
display(ipd.Audio(audio, rate=pipe.unet.sample_rate))AudioPipelineOutput
class diffusers.AudioPipelineOutput
< source >( audios: ndarray )
Output class for audio pipelines.