+hSSKrSSKJrJrJrJrJrJr SSKrSSK J r SSK J r J r \(aSSKJr "SS \ 5rg) N) TYPE_CHECKINGDictListOptionalTupleUnion)register_to_config) BaseGuidancerescale_noise_cfg) BlockStatec D^\rSrSrSrSS/r\SS\S\S\S\S \4 U4S jjj5r SS S S \ \ \ \ \ \\ \ 444S\S 4SjjrSS\R$S\ \R$S\R$4Sjjr\S\4Sj5r\S\4Sj5rS\4SjrSrU=r$)ClassifierFreeGuidancea Classifier-free guidance (CFG): https://huggingface.co/papers/2207.12598 CFG is a technique used to improve generation quality and condition-following in diffusion models. It works by jointly training a model on both conditional and unconditional data, and using a weighted sum of the two during inference. This allows the model to tradeoff between generation quality and sample diversity. The original paper proposes scaling and shifting the conditional distribution based on the difference between conditional and unconditional predictions. [x_pred = x_cond + scale * (x_cond - x_uncond)] Diffusers implemented the scaling and shifting on the unconditional prediction instead based on the [Imagen paper](https://huggingface.co/papers/2205.11487), which is equivalent to what the original paper proposed in theory. [x_pred = x_uncond + scale * (x_cond - x_uncond)] The intution behind the original formulation can be thought of as moving the conditional distribution estimates further away from the unconditional distribution estimates, while the diffusers-native implementation can be thought of as moving the unconditional distribution towards the conditional distribution estimates to get rid of the unconditional predictions (usually negative features like "bad quality, bad anotomy, watermarks", etc.) The `use_original_formulation` argument can be set to `True` to use the original CFG formulation mentioned in the paper. By default, we use the diffusers-native implementation that has been in the codebase for a long time. Args: guidance_scale (`float`, defaults to `7.5`): The scale parameter for classifier-free guidance. Higher values result in stronger conditioning on the text prompt, while lower values allow for more freedom in generation. Higher values may lead to saturation and deterioration of image quality. guidance_rescale (`float`, defaults to `0.0`): The rescale factor applied to the noise predictions. This is used to improve image quality and fix overexposure. Based on Section 3.4 from [Common Diffusion Noise Schedules and Sample Steps are Flawed](https://huggingface.co/papers/2305.08891). use_original_formulation (`bool`, defaults to `False`): Whether to use the original formulation of classifier-free guidance as proposed in the paper. By default, we use the diffusers-native implementation that has been in the codebase for a long time. See [~guiders.classifier_free_guidance.ClassifierFreeGuidance] for more details. start (`float`, defaults to `0.0`): The fraction of the total number of denoising steps after which guidance starts. stop (`float`, defaults to `1.0`): The fraction of the total number of denoising steps after which guidance stops. pred_cond pred_uncondguidance_scaleguidance_rescaleuse_original_formulationstartstopcH>[TU]XE5 XlX lX0lgN)super__init__rrr)selfrrrrr __class__s d/home/james-whalen/.local/lib/python3.13/site-packages/diffusers/guiders/classifier_free_guidance.pyrClassifierFreeGuidance.__init__Gs$ %, 0(@%datar input_fieldsreturncUc URnURS:XaS/OSS/n/n[UR5H6nURX!X5URU5nUR U5 M8 U$)Nr r) _input_fieldsnum_conditionsrange_prepare_batch_input_predictionsappend)rr"r# tuple_indices data_batchesi data_batchs rprepare_inputs%ClassifierFreeGuidance.prepare_inputsVs  --L#22a7aV  t**+A,,\AQSWSjSjklSmnJ    +,r!cSnUR5(dUnO*X- nUR(aUOUnX0RU--nURS:a[ X1UR5nU04$)N)_is_cfg_enabledrrrr )rrrpredshifts rforwardClassifierFreeGuidance.forwardcsn##%%D+E $ = =9;D--55D  3 &$Td6K6KLDRxr!c URS:H$Nr )_count_prepared)rs ris_conditional%ClassifierFreeGuidance.is_conditionalrs##q((r!c>SnUR5(aUS- nU$r:)r4)rr's rr'%ClassifierFreeGuidance.num_conditionsvs&    ! ! a Nr!cUR(dgSnURbb[URUR-5n[URUR-5nX R s=:*=(a U:Os nSnUR (a"[R"URS5nO![R"URS5nU=(a U(+$)NFTr3?) _enabled_num_inference_stepsint_start_stop_steprmathiscloser)ris_within_rangeskip_start_stepskip_stop_stepis_closes rr4&ClassifierFreeGuidance._is_cfg_enabled}s}}  $ $ 0!$++0I0I"IJO d.G.G!GHN-LLnLO  ( (||D$7$7=H||D$7$7=H/x</r!)rrr)g@r3Fr3rAr)__name__ __module__ __qualname____firstlineno____doc__r*r floatboolrrrstrrrrr0torchTensorr7propertyr<rDr'r4__static_attributes__ __classcell__)rs@rrrs2&P&}5!$"%). A A  A#' A  A  A Adh   08c5eTWY\T\oI]C^>^9_0`  l    HU\\ras/ DD 49?q0\q0r!