o Gi @sddlZddlmZddlmZddlmZmZddlm Z ddl m Z m Z m Z er7ddlmmZd Znd ZeeZGd d d e e ZdS) N) UNet1DModel)SchedulerMixin)is_torch_xla_availablelogging) randn_tensor)AudioPipelineOutputDeprecatedPipelineMixinDiffusionPipelineTFcseZdZdZdZdZdedeffdd Ze   dd e d e d e j e e j Bd Bded BdedeeBf ddZZS)DanceDiffusionPipelinea 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.). Parameters: unet ([`UNet1DModel`]): A `UNet1DModel` to denoise the encoded audio. scheduler ([`SchedulerMixin`]): A scheduler to be used in combination with `unet` to denoise the encoded audio latents. Can be one of [`IPNDMScheduler`]. z0.33.1unet schedulercst|j||ddS)N)r r)super__init__register_modules)selfr r __class__p/home/ubuntu/.local/lib/python3.10/site-packages/diffusers/pipelines/dance_diffusion/pipeline_dance_diffusion.pyr5s zDanceDiffusionPipeline.__init__dNT batch_sizenum_inference_steps generatoraudio_length_in_s return_dictreturncCs|dur|jjj|jjj}||jjj}dt|jj}|d|kr4t|dd||jjjdt|}||dkrc||jjj|d|}t |d||jjjd ||jjjd t|}t |j j } ||jjj |f} t|trt||krtd t|d |d t| ||j| d} |jj|| jd|jj| |j_||jjD]} || | j} |j| | | j} trtq| dd !} | ddddd|f} |s| fSt"| dS)a The call function to the pipeline for generation. Args: 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*): A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation deterministic. audio_length_in_s (`float`, *optional*, defaults to `self.unet.config.sample_size/self.unet.config.sample_rate`): The length of the generated audio sample in seconds. return_dict (`bool`, *optional*, defaults to `True`): Whether or not to return a [`~pipelines.AudioPipelineOutput`] instead of a plain tuple. Example: ```py 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 display in google colab import IPython.display as ipd for audio in audios: display(ipd.Audio(audio, rate=pipe.unet.sample_rate)) ``` Returns: [`~pipelines.AudioPipelineOutput`] or `tuple`: If `return_dict` is `True`, [`~pipelines.AudioPipelineOutput`] is returned, otherwise a `tuple` is returned where the first element is a list with the generated audio. Nrrz1 is too small. Make sure it's bigger or equal to .rrz is increased to z; so that it can be handled by the model. It will be cut to z after the denoising process.z/You have passed a list of generators of length z+, but requested an effective batch size of z@. Make sure the batch size matches the length of the generators.)rdevicedtype)r )audios)#r config sample_size sample_ratelen up_blocks ValueErrorintloggerinfonext parametersr! in_channels isinstancelistr_execution_devicer set_timestepsr timestepsto progress_barsamplestep prev_sample XLA_AVAILABLExm mark_stepclampfloatcpunumpyr )rrrrrrr%down_scale_factororiginal_sample_sizer!shapeaudiot model_outputrrr__call__9sV7     zDanceDiffusionPipeline.__call__)rrNNT)__name__ __module__ __qualname____doc___last_supported_versionmodel_cpu_offload_seqrrrtorchno_gradr* Generatorr1r>boolr tuplerG __classcell__rrrrr #s0r )rNmodelsr schedulersrutilsrrutils.torch_utilsrpipeline_utilsr r r torch_xla.core.xla_modelcore xla_modelr;r: get_loggerrHr+r rrrrs