oMj(}bSSKrSSKJrJr SSKrSSKrSSKJrJ r SSK J r J r SSK Jr SSKJrJr SSKJr SS KJrJrJrJr SS KJr SS KJr S S KJr SSKJ r \"5(a SSK!J"s J#r$ Sr%OSr%\RL"\'5r(\"5(aSSK)r)Sr*Sr+Sr,Sr-"SS\\5r.g)N)AnyCallable) AutoTokenizerUMT5EncoderModel)MultiPipelineCallbacksPipelineCallback)WanLoraLoaderMixin)AutoencoderKLWanWanTransformer3DModel)FlowMatchEulerDiscreteScheduler)is_ftfy_availableis_torch_xla_availableloggingreplace_example_docstring) randn_tensor)VideoProcessor)DiffusionPipeline)WanPipelineOutputTFa Examples: ```python >>> import torch >>> from diffusers.utils import export_to_video >>> from diffusers import AutoencoderKLWan, WanPipeline >>> from diffusers.schedulers.scheduling_unipc_multistep import UniPCMultistepScheduler >>> # Available models: Wan-AI/Wan2.1-T2V-14B-Diffusers, Wan-AI/Wan2.1-T2V-1.3B-Diffusers >>> model_id = "Wan-AI/Wan2.1-T2V-14B-Diffusers" >>> vae = AutoencoderKLWan.from_pretrained(model_id, subfolder="vae", torch_dtype=torch.float32) >>> pipe = WanPipeline.from_pretrained(model_id, vae=vae, torch_dtype=torch.bfloat16) >>> flow_shift = 5.0 # 5.0 for 720P, 3.0 for 480P >>> pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config, flow_shift=flow_shift) >>> pipe.to("cuda") >>> prompt = "A cat and a dog baking a cake together in a kitchen. The cat is carefully measuring flour, while the dog is stirring the batter with a wooden spoon. The kitchen is cozy, with sunlight streaming through the window." >>> negative_prompt = "Bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards" >>> output = pipe( ... prompt=prompt, ... negative_prompt=negative_prompt, ... height=720, ... width=1280, ... num_frames=81, ... guidance_scale=5.0, ... ).frames[0] >>> export_to_video(output, "output.mp4", fps=16) ``` c[5(a[R"U5n[R"[R"U55nUR 5$N)rftfyfix_texthtmlunescapestriptexts ڲC:\Users\johnn\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.13_qbz5n2kfra8p0\LocalCache\local-packages\Python313\site-packages\diffusers/pipelines/wan/pipeline_wan.py basic_cleanr"Ns<}}T" ==t, -D ::<cV[R"SSU5nUR5nU$)Nz\s+ )resubrrs r!whitespace_cleanr(Us$ 66&#t $D ::S\\\-S\S\S\R(S-S\R*S-4 SjjrS?S\\\-S\\\-S-S\S\S\R.S-S\R.S-S\S\R(S-S\R*S-4SjjrS@SjrSAS \S!\S"\S#\S$\S\R*S-S\R(S-S%\R4\\R4-S-S&\R.S-S'\R.4S(jjr\S)5r\S*5r\S+5r\S,5r \S-5r!\S.5r"\RF"5\$"\%5SSSSSS/S0SSSSSSS1SSSS&/S24S\\\-S\\\-S"\S#\S$\S3\S4\ S5\ S-S\S-S%\R4\\R4-S-S&\R.S-S\R.S-S\R.S-S6\S-S7\S8\&\\'4S-S9\(\\/S4\)-\*-S-S:\\S\4&S;jj55r+S= boundary_timestep and `transformer_2` handles timesteps < boundary_timestep. If `None`, only `transformer` is used for the entire denoising process. z-text_encoder->transformer->transformer_2->vae)latents prompt_embedsnegative_prompt_embeds transformer transformer_2N tokenizer text_encodervae schedulerboundary_ratioexpand_timestepsc >[T U]5 URUUUUUUS9 URUS9 URUS9 [ USS5(a UR R ROSUl[ USS5(a UR R ROSUl [URS9Ul g)N)r5r4r3r1r6r2)r7)r8r5)vae_scale_factor) super__init__register_modulesregister_to_configgetattrr5configscale_factor_temporalvae_scale_factor_temporalscale_factor_spatialvae_scale_factor_spatialrvideo_processor) selfr3r4r5r6r1r2r7r8 __class__s r!r>WanPipeline.__init__s  %#'   ~> 1ABRYZ^`egkRlRl)N)Nrs&PWX\^ceiPjPj(L(Lpq%-t?\?\]r#rpromptnum_videos_per_promptmax_sequence_lengthdevicedtypecU=(d URnU=(d URRn[U[5(aU/OUnUVs/sHn[ U5PM nn[ U5nURUSUSSSSS9nURURpU RS5RSS9R5n URU RU5U RU55Rn U RXTS9n [X5VV s/sH upmUSU PM n nn [ R""U Vs/sHJn[ R$"XfR'X6R)S5- UR)S55/5PML snSS9n U R*upnU R-SUS5n U R/Xr-US 5n U $s snfs sn nfs snf) N max_lengthTpt)paddingrQ truncationadd_special_tokensreturn_attention_maskreturn_tensorsrr)dimrOrN)_execution_devicer4rO isinstancestrr*lenr3 input_idsattention_maskgtsumlongtolast_hidden_stateziptorchstackcat new_zerossizeshaperepeatview)rHrKrLrMrNrOu batch_size text_inputstext_input_idsmaskseq_lensr/v_seq_lens r!_get_t5_prompt_embeds!WanPipeline._get_t5_prompt_embedss14110**00'44&&+126a,q/62[ nn  *#"&%  +44k6P6P771:>>a>(--/)).*;*;F*CTWWV_Ugg %((u(D +.}+GH+G412A+G H ^k l^kYZUYY;;':VVAY'Fq RS T^k lrs &++ A%,,Q0EqI %**:+MwXZ[ 73"I lsG$G! AG'Tnegative_promptdo_classifier_free_guidancer/r0c 2U=(d URn[U[5(aU/OUnUb [U5n OURSn UcUR UUUUU S9nU(aUcU=(d Sn[U[5(aX/-OUnUb;[ U5[ U5La$[S[ U5S[ U5S35eU [U5:wa![SUS[U5S USU S 3 5eUR UUUUU S9nXV4$) ab Encodes the prompt into text encoder hidden states. Args: prompt (`str` or `list[str]`, *optional*): prompt to be encoded negative_prompt (`str` or `list[str]`, *optional*): The prompt or prompts not to guide the image generation. If not defined, one has to pass `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than `1`). do_classifier_free_guidance (`bool`, *optional*, defaults to `True`): Whether to use classifier free guidance or not. num_videos_per_prompt (`int`, *optional*, defaults to 1): Number of videos that should be generated per prompt. torch device to place the resulting embeddings on prompt_embeds (`torch.Tensor`, *optional*): Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, text embeddings will be generated from `prompt` input argument. negative_prompt_embeds (`torch.Tensor`, *optional*): Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input argument. device: (`torch.device`, *optional*): torch device dtype: (`torch.dtype`, *optional*): torch dtype r)rKrLrMrNrOz?`negative_prompt` should be the same type to `prompt`, but got z != .z`negative_prompt`: z has batch size z, but `prompt`: zT. Please make sure that passed `negative_prompt` matches the batch size of `prompt`.) r[r\r]r^rlrxtype TypeError ValueError) rHrKrzr{rLr/r0rMrNrOrps r! encode_promptWanPipeline.encode_promptskL1411'44&&  VJ&,,Q/J   66&;$7 7M '+A+I-3O@J?\_@`@`j+<# UHoTR;v M g7fr)_callback_tensor_inputs).0krHs r! +WanPipeline.check_inputs..&sF 7Y!-- -7Ysz2`callback_on_step_end_tensor_inputs` has to be in z , but found zCannot forward both `prompt`: z and `prompt_embeds`: z2. Please make sure to only forward one of the two.z'Cannot forward both `negative_prompt`: z and `negative_prompt_embeds`: zeProvide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined.z2`prompt` has to be of type `str` or `list` but is z;`negative_prompt` has to be of type `str` or `list` but is zV`guidance_scale_2` is only supported when the pipeline's `boundary_ratio` is not None.) rallrr\r]listrrBr7) rHrKrzheightwidthr/r0"callback_on_step_end_tensor_inputsguidance_scale_2rs ` r! check_inputsWanPipeline.check_inputss  B;! urzQWX^W__dejdkklmn n - 9#F 7YF C C DTEaEaDbbn|^pH|^vwkokGkGbGpq|^pHoIJ   -";08N}o^00  (-C-O9/9JJikAjBB00 ^ 5w  FC)@)@TZ\`IaIaQRVW]R^Q_`a a  (?C00OUY9Z9ZZ[_`o[pZqrs s ;; % % -2B2Nuv v3O -1pHs E01E0@Qrpnum_channels_latentsrr num_frames generatorr.returnc LU bU RXvS9$US- UR-S-n UUU [U5UR-[U5UR-4n [ U[ 5(a*[ U5U:wa[S[ U5SUS35e[XXvS9n U $)N)rNrOrz/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.)rrNrO) rdrDintrFr\rr^rr) rHrprrrrrOrNrr.num_latent_framesrls r!prepare_latentsWanPipeline.prepare_latentsEs  ::V:9 9'!^0N0NNQRR   K488 8 J$77 7   i & &3y>Z+GA#i.AQR&<'gi  u&Vr#cUR$r_guidance_scalerHs r!guidance_scaleWanPipeline.guidance_scalees###r#c URS:$)N?rrs r!r{'WanPipeline.do_classifier_free_guidanceis##c))r#cUR$r)_num_timestepsrs r! num_timestepsWanPipeline.num_timestepsms"""r#cUR$r)_current_timesteprs r!current_timestepWanPipeline.current_timestepq%%%r#cUR$r) _interruptrs r! interruptWanPipeline.interruptus r#cUR$r)_attention_kwargsrs r!attention_kwargsWanPipeline.attention_kwargsyrr#2g@npinum_inference_stepsrr output_type return_dictrcallback_on_step_endrcV[U[[45(a URnUR UUUUU U UU5 XPR -S:waA[ RSUR S35 XPR -UR -S-n[US5nURb URRROURRRnURUS-nURUS-nUU-U-nUU-U-nUU:wdUU:wa+[ RSUSUSUSUS USUS 3 5 UUpCURRbUcUnXplXlUUlSUlS UlUR(nUb[U[*5(aSnO3Ub![U[,5(a [/U5nOU R0S nUR3UUUR4U U U UUS 9upURbURR6OURR6nU R9U5n U bU R9U5n UR:R=UUS9 UR:R>nURb URRR@OURRR@nURCUU -UUUU[DRFUU U 5 n [DRH"U R0[DRFUS9n[/U5X`R:RJ-- n[/U5Ul&UR:ROS 5 URRb8URRUR:RRP-n OSn URSUS9n![UU5GH=un"n#URV(aMU#UlU bU#U :aURn$Un%OURn$Un%U R9U5n&URRX(aVUS S SS2SSS2SSS24U#-R[5n'U'R]S 5R_U R0S S5n(OU#R_U R0S 5n(U$RaS5 U$"U&U(U US S9S n)SSS5 UR4(a2U$RaS5 U$"U&U(U US S9S n*SSS5 W*U%W)U*- --n)UR:RcW)U#U S S9S n Ub\0n+UHn,[e5U,U+U,'M U"UU"U#U+5n-U-RgSU 5n U-RgSU 5n U-RgSU 5n U"[/U5S- :Xd)U"S-U:a0U"S-UR:RJ-S :XaU!Ri5 [j(dGM([lRn"5 GM@ SSS5 SUlUS:XGdzU R9URpR65n [DRr"URpRRt5RwSURpRRxSSS5R9U RzU R65n.S[DRr"URpRR|5RwSURpRRxSSS5R9U RzU R65- n/U U/- U.-n URpRU S S9S n0URRU0US9n0OU n0UR5 U(dU04$[U0S9$!,(df  GN=f!,(df  GN=f!,(df  GN=f)a The call function to the pipeline for generation. Args: prompt (`str` or `list[str]`, *optional*): The prompt or prompts to guide the image generation. If not defined, pass `prompt_embeds` instead. negative_prompt (`str` or `list[str]`, *optional*): The prompt or prompts to avoid during image generation. If not defined, pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale` < `1`). height (`int`, defaults to `480`): The height in pixels of the generated image. width (`int`, defaults to `832`): The width in pixels of the generated image. num_frames (`int`, defaults to `81`): The number of frames in the generated video. num_inference_steps (`int`, defaults to `50`): The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. guidance_scale (`float`, defaults to `5.0`): Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://huggingface.co/papers/2207.12598). `guidance_scale` is defined as `w` of equation 2. of [Imagen Paper](https://huggingface.co/papers/2205.11487). Guidance scale is enabled by setting `guidance_scale > 1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`, usually at the expense of lower image quality. guidance_scale_2 (`float`, *optional*, defaults to `None`): Guidance scale for the low-noise stage transformer (`transformer_2`). If `None` and the pipeline's `boundary_ratio` is not None, uses the same value as `guidance_scale`. Only used when `transformer_2` and the pipeline's `boundary_ratio` are not None. num_videos_per_prompt (`int`, *optional*, defaults to 1): The number of images to generate per prompt. generator (`torch.Generator` or `list[torch.Generator]`, *optional*): A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation deterministic. latents (`torch.Tensor`, *optional*): Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image generation. Can be used to tweak the same generation with different prompts. If not provided, a latents tensor is generated by sampling using the supplied random `generator`. prompt_embeds (`torch.Tensor`, *optional*): Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the `prompt` input argument. negative_prompt_embeds (`torch.Tensor`, *optional*): Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument. output_type (`str`, *optional*, defaults to `"np"`): The output format of the generated image. Choose between `PIL.Image` or `np.array`. return_dict (`bool`, *optional*, defaults to `True`): Whether or not to return a [`WanPipelineOutput`] instead of a plain tuple. attention_kwargs (`dict`, *optional*): A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under `self.processor` in [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py). callback_on_step_end (`Callable`, `PipelineCallback`, `MultiPipelineCallbacks`, *optional*): A function or a subclass of `PipelineCallback` or `MultiPipelineCallbacks` that is called at the end of each denoising step during the inference. with the following arguments: `callback_on_step_end(self: DiffusionPipeline, step: int, timestep: int, callback_kwargs: Dict)`. `callback_kwargs` will include a list of all tensors as specified by `callback_on_step_end_tensor_inputs`. callback_on_step_end_tensor_inputs (`list`, *optional*): The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the `._callback_tensor_inputs` attribute of your pipeline class. max_sequence_length (`int`, defaults to `512`): The maximum sequence length of the text encoder. If the prompt is longer than this, it will be truncated. If the prompt is shorter, it will be padded to this length. Examples: Returns: [`~WanPipelineOutput`] or `tuple`: If `return_dict` is `True`, [`WanPipelineOutput`] is returned, otherwise a `tuple` is returned where the first element is a list with the generated images and the second element is a list of `bool`s indicating whether the corresponding generated image contains "not-safe-for-work" (nsfw) content. rz(`num_frames - 1` has to be divisible by z!. Rounding to the nearest number.Nrz+`height` and `width` must be multiples of (z, z() for proper patchification. Adjusting (z) -> (z).Fr)rKrzr{rLr/r0rMrN)rNrY)totalrZcond) hidden_statestimestepencoder_hidden_statesrruncond)rr.r/r0latentr)r)frames)Dr\r r tensor_inputsrrDloggerwarningmaxr1rB patch_sizer2rFr7r_guidance_scale_2rrrr[r]rr^rlrr{rOrdr6 set_timesteps timesteps in_channelsrrgfloat32onesorderrset_begin_indexnum_train_timesteps progress_bar enumeraterr8flatten unsqueezeexpand cache_contextsteplocalspopupdate XLA_AVAILABLExm mark_stepr5tensor latents_meanrnz_dimrN latents_stddecoderGpostprocess_videomaybe_free_model_hooksr)1rHrKrzrrrrrrrLrr.r/r0rrrrrrMr h_multiple_of w_multiple_of calc_height calc_widthrNrptransformer_dtyperrrsnum_warmup_stepsboundary_timesteprit current_modelcurrent_guidance_scalelatent_model_inputtemp_tsr noise_pred noise_uncondcallback_kwargsrcallback_outputsrrvideos1 r!__call__WanPipeline.__call__}sB *-=?U,V W W1E1S1S .       " .   66 6! ; NN:4;Y;Y:ZZ{| $'E'EEHfHffijjJQ' +    # # . .##**55  55 1 E 55 1 E  - = m+m; [ EZ$7 NN=m_B}o^$XRwf[MJ%>?P%Q " $$%8$HNN,, +    # # / /##**66  && . .    MM     zz'--u}}VLy>,?..BVBV,VV!)n &&q) ;; % % 1 $ : :T^^=R=R=f=f f  $    %8  9\!),1>>)*&$,5F0F$($4$4M-;*%)$6$6M-=*%,ZZ0A%B";;//#Awqz!SqS#A#+6:CCEG&003::7==;KRPH xx a(89H"008!.&8!).;)9$) "  "J933&44X>'4*<%-2H-=(- (  ( ?".0F*WcJc0d!dJ..--j!WRW-XYZ['3&(O?-3Xa[*@';D!Q'X$.229gFG$4$8$8-$XM-=-A-ABZ\r-s*I**A9I/IqSTuX\XfXfXlXlNlpqNq '') =LLN}-:B"&h&jj0G TXX__99:a..1a8GNNGMM2   TXX__-H-H I N NqRVRZRZRaRaRgRgijlmop q t t !K +l:GHHOOGO?BE((::5k:ZEE ##%8O ..A98?>G: 9sD2D`6_5*`.`r]rrrgrNrOrxTensorrr Generatorrpropertyrr{rrrrno_gradrEXAMPLE_DOC_STRINGdictrrr rr__static_attributes__ __classcell__)rIs@r!r,r,`s<LT)?;596:'+!&^ ^'^ ^ 3 ^ +T1 ^-t3^ ^^^:#'%&#&&*$( 'd3i' #'! '  t# ' {{T! 'X37,0%&-16:#&&*$(O5d3iO5tCy4/O5&* O5 # O5 ||d* O5!& t 3O5!O5 t#O5{{T!O5n#+/+w`%'$(&*DH'+"     {{T! t#??T%//%::TA$ @$$**##&&&& ]]_12#'+/#% #)-,-DH'+-16:"& 26nr9B #&)g/d3ig/tCyg/ g/  g/  g/!g/g/ $,g/ #Tzg/??T%//%::TAg/$g/||d*g/!& t 3g/4Zg/ !g/"sCx.4/#g/$'Sz4'78;KKNddgkk%g/&-1I'g/(!)g/3g/r#r,)/rtypingrrregexr&rg transformersrr callbacksrr loadersr modelsr r schedulersr utilsrrrrutils.torch_utilsrrGrpipeline_utilsrpipeline_outputrtorch_xla.core.xla_modelcore xla_modelrr get_loggerrrrr r"r(r*r,r#r!r!s  8A)=9bb--..))MM   H %@  F /#%7F /r#