o piB3@sddlmZddlmZmZmZddlZddlmZ ddl m Z m Z ddl mZmZmZmZmZmZejjGdd d ZeGd d d eZGd d d ee ZdS)) dataclass)OptionalTupleUnionN) ConfigMixinregister_to_config)CommonSchedulerStateFlaxKarrasDiffusionSchedulersFlaxSchedulerMixinFlaxSchedulerOutputadd_noise_commonget_velocity_commonc @sheZdZUeed<ejed<ejed<ejed<dZee ed<e dedejdejdejfddZ dS) DDIMSchedulerStatecommonfinal_alpha_cumprodinit_noise_sigma timestepsNnum_inference_stepscCs|||||dS)Nrrrr)clsrrrrrrg/home/ubuntu/SoloSpeech/.venv/lib/python3.10/site-packages/diffusers/schedulers/scheduling_ddim_flax.pycreate-s zDDIMSchedulerState.create) __name__ __module__ __qualname__r __annotations__jnpndarrayrrint classmethodrrrrrr#s    rc@seZdZUeed<dS)FlaxDDIMSchedulerOutputstateN)rrrrrrrrrr#=s r#c@seZdZUdZddeDZejed<e ddZ e ddd d d d d d ddej f de dedededeejdededede dedejfddZd;deedefddZ d;d ed!ejd"ee dejfd#d$Z %dFlaxDDIMSchedulerav Denoising diffusion implicit models is a scheduler that extends the denoising procedure introduced in denoising diffusion probabilistic models (DDPMs) with non-Markovian guidance. [`~ConfigMixin`] takes care of storing all config attributes that are passed in the scheduler's `__init__` function, such as `num_train_timesteps`. They can be accessed via `scheduler.config.num_train_timesteps`. [`SchedulerMixin`] provides general loading and saving functionality via the [`SchedulerMixin.save_pretrained`] and [`~SchedulerMixin.from_pretrained`] functions. For more details, see the original paper: https://arxiv.org/abs/2010.02502 Args: num_train_timesteps (`int`): number of diffusion steps used to train the model. beta_start (`float`): the starting `beta` value of inference. beta_end (`float`): the final `beta` value. beta_schedule (`str`): the beta schedule, a mapping from a beta range to a sequence of betas for stepping the model. Choose from `linear`, `scaled_linear`, or `squaredcos_cap_v2`. trained_betas (`jnp.ndarray`, optional): option to pass an array of betas directly to the constructor to bypass `beta_start`, `beta_end` etc. clip_sample (`bool`, default `True`): option to clip predicted sample between for numerical stability. The clip range is determined by `clip_sample_range`. clip_sample_range (`float`, default `1.0`): the maximum magnitude for sample clipping. Valid only when `clip_sample=True`. set_alpha_to_one (`bool`, default `True`): each diffusion step uses the value of alphas product at that step and at the previous one. For the final step there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`, otherwise it uses the value of alpha at step 0. steps_offset (`int`, default `0`): An offset added to the inference steps, as required by some model families. prediction_type (`str`, default `epsilon`): indicates whether the model predicts the noise (epsilon), or the samples. One of `epsilon`, `sample`. `v-prediction` is not supported for this scheduler. dtype (`jnp.dtype`, *optional*, defaults to `jnp.float32`): the `dtype` used for params and computation. cCsg|]}|jqSr)name).0errr iszFlaxDDIMScheduler.dtypecCsdS)NTrselfrrr has_statemszFlaxDDIMScheduler.has_stateig-C6?g{Gz?linearNT?repsilonnum_train_timesteps beta_startbeta_end beta_schedule trained_betas clip_sampleclip_sample_rangeset_alpha_to_one steps_offsetprediction_typec Cs | |_dSNr*) r,r1r2r3r4r5r6r7r8r9r:r*rrr__init__qs zFlaxDDIMScheduler.__init__rreturncCst|dur t|}|jjrtjd|jdn|jd}tjd|jd}td|jj  ddd}t j||||dS)Nr/r<rr) r rconfigr8rarrayr*alphas_cumprodaranger1roundr)r,rrrrrrr create_states  zFlaxDDIMScheduler.create_stater$sampletimestepcCs|S)a Args: state (`PNDMSchedulerState`): the `FlaxPNDMScheduler` state data class instance. sample (`jnp.ndarray`): input sample timestep (`int`, optional): current timestep Returns: `jnp.ndarray`: scaled input sample r)r,r$rFrGrrrscale_model_inputs z#FlaxDDIMScheduler.scale_model_inputrrshapecCs@|jj|}td||ddd|jj}|j||dS)a Sets the discrete timesteps used for the diffusion chain. Supporting function to be run before inference. Args: state (`DDIMSchedulerState`): the `FlaxDDIMScheduler` state data class instance. num_inference_steps (`int`): the number of diffusion steps used when generating samples with a pre-trained model. rNr?)rr)r@r1rrCrDr9replace)r,r$rrI step_ratiorrrr set_timestepss &zFlaxDDIMScheduler.set_timestepsc CsP|jj|}t|dk|jj||j}d|}d|}||d||}|S)Nrr )rrBrwherer) r,r$rG prev_timestep alpha_prod_talpha_prod_t_prev beta_prod_tbeta_prod_t_prevvariancerrr _get_variances zFlaxDDIMScheduler._get_variance model_outputeta return_dictcCsv|jdur td||jj|j}|jj}|j} ||} t|dk||| } d| } |jj dkrA|| d|| d} |}n>|jj dkrV|} || d| | d}n)|jj dkru| d|| d|} | d|| d|}n td |jj d |jj r| |jj |jj } | |||}||d}d| |d d|}| d| |}|s||fSt||d S) a Predict the sample at the previous timestep by reversing the SDE. Core function to propagate the diffusion process from the learned model outputs (most often the predicted noise). Args: state (`DDIMSchedulerState`): the `FlaxDDIMScheduler` state data class instance. model_output (`jnp.ndarray`): direct output from learned diffusion model. timestep (`int`): current discrete timestep in the diffusion chain. sample (`jnp.ndarray`): current instance of sample being created by diffusion process. return_dict (`bool`): option for returning tuple rather than FlaxDDIMSchedulerOutput class Returns: [`FlaxDDIMSchedulerOutput`] or `tuple`: [`FlaxDDIMSchedulerOutput`] if `return_dict` is True, otherwise a `tuple`. When returning a tuple, the first element is the sample tensor. NzaNumber of inference steps is 'None', you need to run 'set_timesteps' after creating the schedulerrr r0g?rF v_predictionzprediction_type given as z6 must be one of `epsilon`, `sample`, or `v_prediction`r) prev_sampler$)r ValueErrorr@r1rrBrrrMr:r6clipr7rTr#)r,r$rVrGrFrWrXrNrBrrOrPrQpred_original_sample pred_epsilonrS std_dev_tpred_sample_directionrZrrrstepsB      zFlaxDDIMScheduler.steporiginal_samplesnoisercCt|j|||Sr;)rr)r,r$rbrcrrrr add_noise'zFlaxDDIMScheduler.add_noisecCrdr;)rr)r,r$rFrcrrrr get_velocity0rfzFlaxDDIMScheduler.get_velocitycCs|jjSr;)r@r1r+rrr__len__9szFlaxDDIMScheduler.__len__r;)r)rUT)!rrr__doc__r _compatiblesrr*rpropertyr-rfloat32r!floatstrrr boolr=r rrErHrrLrTrr#rarergrhrrrrr%Bs &          ]   r%) dataclassesrtypingrrrflax jax.numpynumpyrconfiguration_utilsrrscheduling_utils_flaxr r r r rrstructrr#r%rrrrs