i@m $ddlZddlmZddlmZmZmZmZmZddl Z ddl Z ddl m Z mZddlmZddlmZdd lmZmZeGd d eZ dd ed ededde j2fdZde j2de j2fdZGddee Zy)N) dataclass)ListLiteralOptionalTupleUnion) ConfigMixinregister_to_config) BaseOutput) randn_tensor)KarrasDiffusionSchedulersSchedulerMixincXeZdZUdZej ed<dZeej ed<y)DDIMSchedulerOutputaq Output class for the scheduler's `step` function output. Args: prev_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images): Computed sample `(x_{t-1})` of previous timestep. `prev_sample` should be used as next model input in the denoising loop. pred_original_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images): The predicted denoised sample `(x_{0})` based on the model output from the current timestep. `pred_original_sample` can be used to preview progress or for guidance. prev_sampleNpred_original_sample) __name__ __module__ __qualname____doc__torchTensor__annotations__rrn/home/obispo/Crisostomo_bridge/mision_env/lib/python3.12/site-packages/diffusers/schedulers/scheduling_ddim.pyrrs' 37(5<<07rrnum_diffusion_timestepsmax_betaalpha_transform_type)cosineexpreturnc $|dk(rd}n|dk(rd}ntd|g}t|D]<}||z }|dz|z }|jtd||||z z |>t j |tj S)aB Create a beta schedule that discretizes the given alpha_t_bar function, which defines the cumulative product of (1-beta) over time from t = [0,1]. Contains a function alpha_bar that takes an argument t and transforms it to the cumulative product of (1-beta) up to that part of the diffusion process. Args: num_diffusion_timesteps (`int`): The number of betas to produce. max_beta (`float`, defaults to `0.999`): The maximum beta to use; use values lower than 1 to avoid numerical instability. alpha_transform_type (`"cosine"` or `"exp"`, defaults to `"cosine"`): The type of noise schedule for `alpha_bar`. Choose from `cosine` or `exp`. Returns: `torch.Tensor`: The betas used by the scheduler to step the model outputs. r"cftj|dzdz tjzdz dzS)NgMb?gT㥛 ?r )mathcospits r alpha_bar_fnz)betas_for_alpha_bar..alpha_bar_fnMs-88QY%/$''9A=>!C Crr#c2tj|dzS)Ng()r'r#r*s rr,z)betas_for_alpha_bar..alpha_bar_fnRs88AI& &rz"Unsupported alpha_transform_type: rdtype) ValueErrorrangeappendminrtensorfloat32)rr r!r,betasit1t2s rbetas_for_alpha_barr:3s0x' D  & '=>R=STUU E * +M ( (!e. . S\"- R0@@@(KLM <<U]] 33rr6c(d|z }tj|d}|j}|dj}|dj}||z}||||z z z}|dz}|dd|ddz }tj|dd|g}d|z }|S)a: Rescales betas to have zero terminal SNR Based on https://huggingface.co/papers/2305.08891 (Algorithm 1) Args: betas (`torch.Tensor`): The betas that the scheduler is being initialized with. Returns: `torch.Tensor`: Rescaled betas with zero terminal SNR. ?rdimr rN)rcumprodsqrtclonecat)r6alphasalphas_cumprodalphas_bar_sqrtalphas_bar_sqrt_0alphas_bar_sqrt_T alphas_bars rrescale_zero_terminal_snrrJ`s5[F]]6q1N$))+O(*002'+113((O(,=@Q,QRRO!!#J ^j"o -F YY 1Q0 1F JE Lrc TeZdZdZeDcgc]}|j c}}ZdZe d.de de de de dd e e ejee fd ed ed e d e ddede de de de ddefdZd/dej(de e dej(fdZde de dej(fdZdej(dej(fdZd/de de eej2fddfd Z d0d!ej(de dej(d"e d#ed$e ej6d%e ej(d&ede eeffd'Zd(ej(d)ej(d*ej>dej(fd+Z dej(d)ej(d*ej>dej(fd,Z!de fd-Z"ycc}}w)1 DDIMSchedulera `DDIMScheduler` extends the denoising procedure introduced in denoising diffusion probabilistic models (DDPMs) with non-Markovian guidance. This model inherits from [`SchedulerMixin`] and [`ConfigMixin`]. Check the superclass documentation for the generic methods the library implements for all schedulers such as loading and saving. Args: num_train_timesteps (`int`, defaults to 1000): The number of diffusion steps to train the model. beta_start (`float`, defaults to 0.0001): The starting `beta` value of inference. beta_end (`float`, defaults to 0.02): The final `beta` value. beta_schedule (`Literal["linear", "scaled_linear", "squaredcos_cap_v2"]`, defaults to `"linear"`): The beta schedule, a mapping from a beta range to a sequence of betas for stepping the model. Must be one of `"linear"`, `"scaled_linear"`, or `"squaredcos_cap_v2"`. trained_betas (`np.ndarray`, *optional*): Pass an array of betas directly to the constructor to bypass `beta_start` and `beta_end`. clip_sample (`bool`, defaults to `True`): Clip the predicted sample for numerical stability. clip_sample_range (`float`, defaults to 1.0): The maximum magnitude for sample clipping. Valid only when `clip_sample=True`. set_alpha_to_one (`bool`, defaults to `True`): Each diffusion step uses the alphas product value 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 alpha value at step 0. steps_offset (`int`, defaults to 0): An offset added to the inference steps, as required by some model families. prediction_type (`Literal["epsilon", "sample", "v_prediction"]`, defaults to `"epsilon"`): Prediction type of the scheduler function. Must be one of `"epsilon"` (predicts the noise of the diffusion process), `"sample"` (directly predicts the noisy sample), or `"v_prediction"` (see section 2.4 of [Imagen Video](https://huggingface.co/papers/2210.02303) paper). thresholding (`bool`, defaults to `False`): Whether to use the "dynamic thresholding" method. This is unsuitable for latent-space diffusion models such as Stable Diffusion. dynamic_thresholding_ratio (`float`, defaults to 0.995): The ratio for the dynamic thresholding method. Valid only when `thresholding=True`. sample_max_value (`float`, defaults to 1.0): The threshold value for dynamic thresholding. Valid only when `thresholding=True`. timestep_spacing (`Literal["leading", "trailing", "linspace"]`, defaults to `"leading"`): The way the timesteps should be scaled. Must be one of `"leading"`, `"trailing"`, or `"linspace"`. Refer to Table 2 of the [Common Diffusion Noise Schedules and Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information. rescale_betas_zero_snr (`bool`, defaults to `False`): Whether to rescale the betas to have zero terminal SNR. This enables the model to generate very bright and dark samples instead of limiting it to samples with medium brightness. Loosely related to [`--offset_noise`](https://github.com/huggingface/diffusers/blob/74fd735eb073eb1d774b1ab4154a0876eb82f055/examples/dreambooth/train_dreambooth.py#L506). rNnum_train_timesteps beta_startbeta_end beta_schedule)linear scaled_linearsquaredcos_cap_v2 trained_betas clip_sampleset_alpha_to_one steps_offsetprediction_type)epsilonsample v_prediction thresholdingdynamic_thresholding_ratioclip_sample_rangesample_max_valuetimestep_spacing)leadingtrailinglinspacerescale_betas_zero_snrct|+tj|tj|_n|dk(r-tj|||tj|_nk|dk(r6tj|dz|dz|tjdz|_n0|dk(rt ||_nt |d|j|rt|j|_d|jz |_ tj|jd |_ |rtjdn|jd |_ d|_ d|_tjt!j"d |ddd j%j't j(|_y) Nr.rQrR?r rSz is not implemented for r<rr=r?)rr4r5r6rcr:NotImplementedError __class__rJrDr@rEfinal_alpha_cumprodinit_noise_sigmanum_inference_steps from_numpynparangecopyastypeint64 timesteps)selfrMrNrOrPrTrUrVrWrXr\r]r^r_r`rds r__init__zDDIMScheduler.__init__se&  $m5==IDJ h & H>QY^YfYfgDJ o - C3H[chcpcpquvvDJ 1 1,-@ADJ%7OPTP^P^O_&`a a "24::>DJDJJ& #mmDKKQ? 9I5<<#4dNaNabcNd !$$( ))"))A7J*KDbD*Q*V*V*X*_*_`b`h`h*ijrrZtimestepr$c|S)a Ensures interchangeability with schedulers that need to scale the denoising model input depending on the current timestep. Args: sample (`torch.Tensor`): The input sample. timestep (`int`, *optional*): The current timestep in the diffusion chain. Returns: `torch.Tensor`: A scaled input sample. r)rsrZrus rscale_model_inputzDDIMScheduler.scale_model_inputs  r prev_timestepc|j|}|dk\r|j|n |j}d|z }d|z }||z d||z z z}|S)a Computes the variance of the noise added at a given diffusion step. For a given `timestep` and its previous step, this method calculates the variance as defined in DDIM/DDPM literature: var_t = (beta_prod_t_prev / beta_prod_t) * (1 - alpha_prod_t / alpha_prod_t_prev) where alpha_prod and beta_prod are cumulative products of alphas and betas, respectively. Args: timestep (`int`): The current timestep in the diffusion process. prev_timestep (`int`): The previous timestep in the diffusion process. If negative, uses `final_alpha_cumprod`. Returns: `torch.Tensor`: The variance for the current timestep. rr)rEri)rsrurx alpha_prod_talpha_prod_t_prev beta_prod_tbeta_prod_t_prevvariances r _get_variancezDDIMScheduler._get_variancesk&**84 BOSTBTD// >Z^ZrZr,& 00${2q 1, then we threshold xt0 to the range [-s, s] and then divide by s. Dynamic thresholding pushes saturated pixels (those near -1 and 1) inwards, thereby actively preventing pixels from saturation at each step. We find that dynamic thresholding results in significantly better photorealism as well as better image-text alignment, especially when using very large guidance weights." https://huggingface.co/papers/2205.11487 Args: sample (`torch.Tensor`): The predicted sample to be thresholded. Returns: `torch.Tensor`: The thresholded sample. rr=)r3max)r/shaperr5float64floatreshapermprodabsquantileconfigr]clampr_ unsqueezeto)rsrZr/ batch_sizechannelsremaining_dims abs_sampless r_threshold_samplezDDIMScheduler._threshold_samples( 06 - H~  6 6\\^F Hrww~7N,NOZZ\ NN:t{{'M'MST U KK 1$++66  KKNVaR+a/ HF~F5! rrkdevicec ||jjkDr=td|d|jjd|jjd||_|jjdk(rot j d|jjdz |jddd jjt j}nn|jjd k(r|jj|jz}t jd||zjddd jjt j}||jjz }n|jjd k(r|jj|jz }t jt j|jjd| jt j}|dz}n"t|jjd tj|j||_y) a Sets the discrete timesteps used for the diffusion chain (to be run before inference). Args: num_inference_steps (`int`): The number of diffusion steps used when generating samples with a pre-trained model. device (`Union[str, torch.device]`, *optional*): The device to use for the timesteps. Raises: ValueError: If `num_inference_steps` is larger than `self.config.num_train_timesteps`. z`num_inference_steps`: z6 cannot be larger than `self.config.train_timesteps`: zG as the unet model trained with this scheduler can only handle maximal z timesteps.rcrrNr?rarbzM is not supported. Please make sure to choose one of 'leading' or 'trailing'.)rrMr0rkr`rmrcroundrorprqrnrWrrlrrr)rsrkrrr step_ratios r set_timestepszDDIMScheduler.set_timestepsGs !@!@ @)*=)>?KK3345 KK;;>BDWX2!  [[ ) )Y 688DZ^ZrZr,&  ;; & &) 3$*[S-AL-P$PT`ehTi#i 'L [[ ( (H 4#/ "\c%:=Q%QQU`ehUiiL [[ ( (N :$0#$5#?;PSCSWcBc#c (#-=cAQU[@[+DKK,G,G+HI""  ;; # ##'#9#9:N#O [[ $ $#7#=#=... 0M0M$ %%h >(s++ #"\c%:=Q%QQU`ehUiiL"#%6!6A!E3 OR^ ^(C03GGJ__ 7)i.C 6 %!- &&)LDWDW_k_q_q"!>1H%0K$  #{Qeffroriginal_samplesnoiserrc|jj|j|_|jj|j}|j|j}||dz}|j }t |j t |j kr=|jd}t |j t |j kr=d||z dz}|j }t |j t |j kr=|jd}t |j t |j kr=||z||zz}|S)a2 Add noise to the original samples according to the noise magnitude at each timestep (this is the forward diffusion process). Args: original_samples (`torch.Tensor`): The original samples to which noise will be added. noise (`torch.Tensor`): The noise to add to the samples. timesteps (`torch.IntTensor`): The timesteps indicating the noise level for each sample. Returns: `torch.Tensor`: The noisy samples. rr.rfr?rrErrr/flattenlenrr)rsrrrrrEsqrt_alpha_prodsqrt_one_minus_alpha_prod noisy_sampless r add_noisezDDIMScheduler.add_noisesa2#1144>qAFF>L E$("QYBF !%HQ",1#&"%GP',!1k 1k1k 1k MN 1k  bjj$u+&= >? 1k1k1k1k!!DE1k1k%*1k!1k 1k""CD1k !%!1k1kf  Y^YeYe"c#%,,:) ))V0@0@eCDU>V0@bf0@n)./315 BgllBgBg Bg  Bg #' BgEOO,Bg!.BgBg "E) *BgJ(,,(||(?? (  (V 5<<  QVQ`Q` ejeqeq D//i ?sF$rL)g+?r")r' dataclassesrtypingrrrrrnumpyrmrconfiguration_utilsr r utilsr utils.torch_utilsr scheduling_utilsrrrrrrr:rJrLrrrrs$ !88 A,G 8*8 8(5=*4 *4*4"/2*4 \\ *4Z!U\\!ell!HH/NKH/r