biVddlmZmZmZmZmZmZddlZddl Z ddl m Z m Z m Z mZddlmZmZmZddlmZddlmZddlmZdd lmZmZmZmZmZmZdd l m!Z!d d l"m#Z#m$Z$m%Z%d d l&m'Z'd dl(m)Z)ddl*m+Z+er ddl,m-cm.Z/dZ0ndZ0ejbe2Z3dZ4Gdde)eeeZ5y))AnyCallableDictListOptionalUnionN) CLIPTextModel CLIPTokenizerT5EncoderModelT5TokenizerFast)FluxLoraLoaderMixinFromSingleFileMixinTextualInversionLoaderMixin) AutoencoderKL)FluxTransformer2DModel)FlowMatchEulerDiscreteScheduler)USE_PEFT_BACKENDis_torch_xla_availableloggingreplace_example_docstringscale_lora_layersunscale_lora_layers) randn_tensor)calculate_shiftretrieve_latentsretrieve_timesteps)FluxPipelineOutput)DiffusionPipeline)VisualClozeProcessorTFaD Examples: ```python >>> import torch >>> from diffusers import VisualClozeGenerationPipeline, FluxFillPipeline as VisualClozeUpsamplingPipeline >>> from diffusers.utils import load_image >>> from PIL import Image >>> image_paths = [ ... # in-context examples ... [ ... load_image( ... "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/visualcloze/visualcloze_mask2image_incontext-example-1_mask.jpg" ... ), ... load_image( ... "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/visualcloze/visualcloze_mask2image_incontext-example-1_image.jpg" ... ), ... ], ... # query with the target image ... [ ... load_image( ... "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/visualcloze/visualcloze_mask2image_query_mask.jpg" ... ), ... None, # No image needed for the target image ... ], ... ] >>> task_prompt = "In each row, a logical task is demonstrated to achieve [IMAGE2] an aesthetically pleasing photograph based on [IMAGE1] sam 2-generated masks with rich color coding." >>> content_prompt = "Majestic photo of a golden eagle perched on a rocky outcrop in a mountainous landscape. The eagle is positioned in the right foreground, facing left, with its sharp beak and keen eyes prominently visible. Its plumage is a mix of dark brown and golden hues, with intricate feather details. The background features a soft-focus view of snow-capped mountains under a cloudy sky, creating a serene and grandiose atmosphere. The foreground includes rugged rocks and patches of green moss. Photorealistic, medium depth of field, soft natural lighting, cool color palette, high contrast, sharp focus on the eagle, blurred background, tranquil, majestic, wildlife photography." >>> pipe = VisualClozeGenerationPipeline.from_pretrained( ... "VisualCloze/VisualClozePipeline-384", resolution=384, torch_dtype=torch.bfloat16 ... ) >>> pipe.to("cuda") >>> image = pipe( ... task_prompt=task_prompt, ... content_prompt=content_prompt, ... image=image_paths, ... guidance_scale=30, ... num_inference_steps=30, ... max_sequence_length=512, ... generator=torch.Generator("cpu").manual_seed(0), ... ).images[0][0] >>> # optional, upsampling the generated image >>> pipe_upsample = VisualClozeUpsamplingPipeline.from_pipe(pipe) >>> pipe_upsample.to("cuda") >>> mask_image = Image.new("RGB", image.size, (255, 255, 255)) >>> image = pipe_upsample( ... image=image, ... mask_image=mask_image, ... prompt=content_prompt, ... width=1344, ... height=768, ... strength=0.4, ... guidance_scale=30, ... num_inference_steps=30, ... max_sequence_length=512, ... generator=torch.Generator("cpu").manual_seed(0), ... ).images[0] >>> image.save("visualcloze.png") ``` c%beZdZdZdZgZddgZ d=dedede de d e d e d e d effd Z d>deeeefdededeej*deej,f dZ d?deeeefdedeej*fdZ d@deeeefdeeeefdeeeefdeej*dedeej2deej2dedeefdZdej8dej:fd Zd!Z dAd"Z e!d#Z"e!d$Z#e!d%Z$d&Z%d'Z&d(Z'd)Z(d*Z)d+Z*e+d,Z,e+d-Z-e+d.Z.e+d/Z/ej`e1e2dddd0dd1dddddd2d3dddgdfdeeeefdeeeefdeej2d4ed5eeed6edeedeeej:eej:fdeej2deej2deej2d7eed8e3d9ee4ee5fd:ee6eee4gdfd;eedef"d<Z7xZ8S)BVisualClozeGenerationPipelinea The VisualCloze pipeline for image generation with visual context. Reference: https://github.com/lzyhha/VisualCloze/tree/main This pipeline is designed to generate images based on visual in-context examples. Args: transformer ([`FluxTransformer2DModel`]): Conditional Transformer (MMDiT) architecture to denoise the encoded image latents. scheduler ([`FlowMatchEulerDiscreteScheduler`]): A scheduler to be used in combination with `transformer` to denoise the encoded image latents. vae ([`AutoencoderKL`]): Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations. text_encoder ([`CLIPTextModel`]): [CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant. text_encoder_2 ([`T5EncoderModel`]): [T5](https://huggingface.co/docs/transformers/en/model_doc/t5#transformers.T5EncoderModel), specifically the [google/t5-v1_1-xxl](https://huggingface.co/google/t5-v1_1-xxl) variant. tokenizer (`CLIPTokenizer`): Tokenizer of class [CLIPTokenizer](https://huggingface.co/docs/transformers/en/model_doc/clip#transformers.CLIPTokenizer). tokenizer_2 (`T5TokenizerFast`): Second Tokenizer of class [T5TokenizerFast](https://huggingface.co/docs/transformers/en/model_doc/t5#transformers.T5TokenizerFast). resolution (`int`, *optional*, defaults to 384): The resolution of each image when concatenating images from the query and in-context examples. z.text_encoder->text_encoder_2->transformer->vaelatents prompt_embeds schedulervae text_encoder tokenizertext_encoder_2 tokenizer_2 transformer resolutionc t ||j|||||||||_t |ddr/dt |j jjdz znd|_ t |ddr |j jjnd|_ t|jdz|j||_ t|dr"|j|jjnd |_d |_y) N)r(r)r+r*r,r-r'r(rr!)vae_scale_factorvae_latent_channelsr.r*M)super__init__register_modulesr.getattrlenr(configblock_out_channelsr2latent_channelsr"image_processorhasattrr*model_max_lengthtokenizer_max_lengthdefault_sample_size) selfr'r(r)r*r+r,r-r. __class__s z/home/cdr/jupyterlab/.venv/lib/python3.12/site-packages/diffusers/pipelines/visualcloze/pipeline_visualcloze_generation.pyr7z&VisualClozeGenerationPipeline.__init__s  %)##  %V]^bdikoVpc$((//*L*L&MPQ&Q RvwCJ$PUW[B\txx>>bd3!22Q6DL`L`mw 07t[/IdnnNhDNN + +np !$' Nr!promptnum_images_per_promptmax_sequence_lengthdevicedtypec |xs |j}|xs|jj}t|tr|gn|}t |}t|t r|j||j}|j|d|dddd}|j}|j|ddj} | jd|jdk\rbtj|| sL|jj| dd|jd z df} tj!d |d | |j#|j%|d d } |j"j}| j%||} | j\} } } | j'd |d } | j)||z| d} | S)N max_lengthTFpt)paddingrN truncation return_lengthreturn_overflowing_tokensreturn_tensorslongestrPrTr!zXThe following part of your input was truncated because `max_sequence_length` is set to tokens: output_hidden_statesrrLrK)_execution_devicer)rL isinstancestrr:rmaybe_convert_promptr, input_idsshapetorchequal batch_decoderAloggerwarningr+torepeatview)rCrHrIrJrKrL batch_size text_inputstext_input_idsuntruncated_ids removed_textr&_seq_lens rE_get_t5_prompt_embedsz3VisualClozeGenerationPipeline._get_t5_prompt_embedss14110**00'4&&[ d7 8..vt7G7GHF&&  *&+' %..**69UY*Zdd   $(<(>66q$JcJcfgJgjlJlGl7mnL NN--.i ~G )).*;*;F*CZ_)` &33 %((t/@/@/F/Fv(V &,,Q0EF %**:8M+MrR rF layout_prompt task_promptcontent_promptpooled_prompt_embeds lora_scalec |xs |j}| gt|trW| |_|jt rt |j| |jt rt |j| t|tr |g}|g}|g}d d} tt|D cgc]7} | || d| || d| || dj9} } |j| ||}|j| |||}|j,t|trt rt|j| |j,t|trt rt|j| |j|jjn|j j} t#j$|j&ddj)|| }|||fScc} w) a Args: layout_prompt (`str` or `List[str]`, *optional*): The prompt or prompts to define the number of in-context examples and the number of images involved in the task. task_prompt (`str` or `List[str]`, *optional*): The prompt or prompts to define the task intention. content_prompt (`str` or `List[str]`, *optional*): The prompt or prompts to define the content or caption of the target image to be generated. device: (`torch.device`): torch device num_images_per_prompt (`int`): number of images that should be generated per prompt prompt_embeds (`torch.FloatTensor`, *optional*): Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, text embeddings will be generated from `prompt` input argument. pooled_prompt_embeds (`torch.FloatTensor`, *optional*): Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, pooled text embeddings will be generated from `prompt` input argument. lora_scale (`float`, *optional*): A lora scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded. c| |rd|S|Sy)Nz(The last image of the last row depicts: )rHcontents rE _preprocessz@VisualClozeGenerationPipeline.encode_prompt.._preprocessTs!!NUA&Ja[aarF T)r~)rHrKrI)rHrIrJrKr!r rKrL)F)r\r]r _lora_scaler)rrr+r^ranger:striprtrqrrLr-rbzerosrarg)rCrurvrwrKrIr&rxrJryrirHrLtext_idss rE encode_promptz+VisualClozeGenerationPipeline.encode_promptsF1411  !j7J&K)D   ,1A!$"3"3Z@"".3C!$"5"5zB mS )*OM&-K,-N 3}-. =+,-Q{;q>/J.K1[YghiYjtxMyLz { B B D   $;;"7 <  22"7 3 3     ($ 349I#D$5$5zB    *$ 349I#D$7$7D+/+<+<+H!!''dN^N^NdNd;;}2215q9<>#4 5 NN * *7T^^5I5I+I J-777rFc ~|Ltfd|Ds8tdjd|Dcgc]}|jvs|c}|| | td|| | td| td|-t|tt fstdt ||-t|tt fstdt |t|t st|t rt|t rt|t s#td t |d t |t|t|k7r td |D]} t| t rt| d t s td t| D chc] } t| c} dk7r tdtd| dDs td| ddD]} td| Dstd| | td||dkDrtd|yycc}wcc} w)Nc3:K|]}|jvywN)_callback_tensor_inputs).0krCs rE z=VisualClozeGenerationPipeline.check_inputs..s#F 23A-- -F sz2`callback_on_step_end_tensor_inputs` has to be in z , but found zOCannot provide both text `task_prompt` + `content_prompt` and `prompt_embeds`. zVMust provide either `task_prompt` + `content_prompt` or pre-computed `prompt_embeds`. z`task_prompt` is missing.z'`task_prompt` must be str or list, got z*`content_prompt` must be str or list, got z[`task_prompt` and `content_prompt` must both be lists, or both be of type str or None, got z and zP`task_prompt` and `content_prompt` must have the same length whe they are lists.rz7Each sample in the batch must have a 2D list of images.r!zKEach in-context example and query should contain the same number of images.c3$K|]}|du ywrr}rimgs rErz=VisualClozeGenerationPipeline.check_inputs..s=33$;=rWzGThere are no targets in the query, which should be represented as None.c3$K|]}|du ywrr}rs rErz=VisualClozeGenerationPipeline.check_inputs..s633$;6rz*Images are missing in in-context examples.zIf `prompt_embeds` are provided, `pooled_prompt_embeds` also have to be passed. Make sure to generate `pooled_prompt_embeds` from the same text encoder that was used to generate `prompt_embeds`.rGz+max_sequence_length cannot exceed 512, got ) all ValueErrorrr]r^rtyper:any) rCrrvrwr&rx"callback_on_step_end_tensor_inputsrJrsamplerows ` rE check_inputsz*VisualClozeGenerationPipeline.check_inputss . 9#F 7YF C DTEaEaDbbn|^pHvwbckokGkGbGpqpHoIJ   #~'A}G`no o  >#9m>Suv v  89 9  ":kC;+OFtKGXFYZ[ [  %j#t.UI$~J^I_`a a k4 (J~t,Lk40 >SW8X  ,-U43G2HJ>"c+&66 !stt W!&$/z&)T7R$%^__F3SC349$%rss=&*==$%noo!#2;WC6#66()UVVW W  $)=)EU   */BS/HJK^J_`a a0I *_pH@4sH5H5H: cg}t|dD]\}}|jd}|j\}}} ||zdz} | |zdz} tj| | d||} || d<tj | ||dddf| d<tj | ||dddf| d <| j d d} |j| tj|d S) Nr!)startrrr r).r).r!).rrWr) enumeratesqueezerarbrarangereshapeappendr) rr2rKrLlatent_image_idsidxrchannelsheightwidth num_patches_h num_patches_w patch_idss rE_prepare_latent_image_idsz7VisualClozeGenerationPipeline._prepare_latent_image_idss!%q1 /HC++a.C&)ii #Hfe"&66!;M!%55:M M=!FZ_`I #If  % ]6QV WXY[_X_ `If  % ]6QV WX\^_X_ `If !))"a0I  # #I . /yy)q11rFc|j|||dzd|dzd}|jdddddd}|j||dz|dzz|dz}|S)Nrrr!r )ripermuter)r%rjnum_channels_latentsrrs rE _pack_latentsz+VisualClozeGenerationPipeline._pack_latentssl,,z+?1aQVZ[Q[]^_//!Q1a3//*v{uz.JL`cdLderFc |j\}}}d}g}tt|D]}||} | dd|z} t| D cgc]} | d c} |z} || | zdzz} |dd|| f}|j || dz| dz|dzdd}|j dddddd}|j ||dz| | }|j|| }|Scc} w)Nrr!rrr r)rarr:sumrirrr)r%sizesr2rj num_patchesrrunpacked_latentsrcur_sizersizerend cur_latentss rE_unpack_latentsz-VisualClozeGenerationPipeline._unpack_latentss,3MM) Ks5z" AQxHa[^'77FX6Ta67;KKE6E>a//C!!U3Y,/K%**:v{EQJPX\]P]_`bcdK%--aAq!Q?K%--j(u:MvW\]K  # #K 0E   7s C c8|jjy)z Enable sliced VAE decoding. When this option is enabled, the VAE will split the input tensor in slices to compute decoding in several steps. This is useful to save some memory and allow larger batch sizes. N)r(enable_slicingrCs rEenable_vae_slicingz0VisualClozeGenerationPipeline.enable_vae_slicing  !rFc8|jjy)z Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to computing decoding in one step. N)r(disable_slicingrs rEdisable_vae_slicingz1VisualClozeGenerationPipeline.disable_vae_slicings   "rFc8|jjy)a Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow processing larger images. N)r( enable_tilingrs rEenable_vae_tilingz/VisualClozeGenerationPipeline.enable_vae_tilings  rFc8|jjy)z Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to computing decoding in one step. N)r(disable_tilingrs rEdisable_vae_tilingz0VisualClozeGenerationPipeline.disable_vae_tiling rrFc @|Dcgc]*}tj|dj||,}}|Dcgc]*}tj|dj||,}}|j||||} |Dcgc]}|j ||} }| Dcgc]}|j } }t t| D]} | | jdd\} }}|j| | d| ||| | <|j| | d| ||| | <|| jdd\} }}|| jd| ||z|||z||| <|| jdddddd || <|| jd| |dzz||z||z|| <|j|| d| |dzz||z||z|| <tj| d} tj| d} tj|d}| | || fScc}wcc}wcc}wcc}w) z6Helper function to prepare latents for a single batch.r rrr!Nrrrr) rbrrgrrclonerr:rarrirr)rCrmaskgenr2rKrLrmr image_latentmasked_image_latentrrrrs rE_prepare_latentsz.VisualClozeGenerationPipeline._prepare_latents'sRWW#3A&))u)EWWLPQq !#&&fE&BQQ 99%AQSY[`aEJJS..sC8J J6BCssyy{CCs<() A2>q/2G2G2K / &%"00a!EY[achiLO%)%7%78KA8NPQSgioqv%w  "37q'--2C / &%1gll$** )) DG1gooaAq!Q7DG1goo$(8!(;<**)) DG ((Q$(8!(;<**)) DG/ @yy15 #ii(;Cyy1%0$8HHH]XQ KCs/H /HH"Hc t|tr)t||k7rtdt|d|dg} g} g} g} t t|D]} |j || || t|t jr|n|| |||\}}}}| j|| j|| j|| j|t j| d} t j| d} t j| d} || jdkDr|| jdzdk(rd|| jdz}t j| g|zd} t j| g|zd} t j| g|zd} ntd| jdd|dt| j||| }|jj| ||j| }t j| | fd j| } || | dfS) Nz/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.rrzCannot expand batch size from z to z.. Batch sizes must be multiples of each other.)rrKrL)rLrW)r]rr:rrrrb Generatorrrrarr' scale_noiserg)rC input_image input_masktimesteprjrLrKrr2masked_image_latentsrmasksrr _image_latent_masked_image_latent_mask_latent_image_idsadditional_image_per_promptnoisesr%s rEprepare_latentsz-VisualClozeGenerationPipeline.prepare_latentsZsw i &3y>Z+GA#i.AQR&<'gi  " s;'( 7ALPLaLaA1 ' 5??C ST  M IM/8I ! ' '(< =   / LL   # #$5 6 7 %yy)=1E -Q7  %Q' ,2215 5066q99Q>.8)FBORRY^R_,.>q.AAArFc|jSr)_guidance_scalers rEguidance_scalez,VisualClozeGenerationPipeline.guidance_scales###rFc|jSr)_joint_attention_kwargsrs rEjoint_attention_kwargsz4VisualClozeGenerationPipeline.joint_attention_kwargss+++rFc|jSr)_num_timestepsrs rE num_timestepsz+VisualClozeGenerationPipeline.num_timestepss"""rFc|jSr) _interruptrs rE interruptz'VisualClozeGenerationPipeline.interrupts rF2g>@pilTrsigmasr output_type return_dictrcallback_on_step_endrc ~ j|||| | |||_|_d_jj |||j }|dt|dtrd}n&|d!t|dtrt|d}j}jjjddnd}j|d|d|d | | |||| \} } }tfd |d d D}t!|j"j$jddj"j$jddj"j$jddj"j$jdd}|t'j(dd|z |n|}t+j"||||\}}j-|d|\}}|ddj/|z}j1|d|d|||z| j2||j \} }}t5t||j"j6zz d }t|_j:j$j<rGt?j@dg||t>jB}|jE| jFd }nd}jI|5}tK|D]\} }!jLr|!jE| jFd jO| j2}"t?jP| |fd}#j;|#|"dz || | ||jd d }$| j2}%j"jS|$|!| d d } | j2|%k7r9t>jTjVjYr| jO|%} |Hi}&|D]}'t[|'|&|'<|| |!|&}(|(j]d!| } |(j]d"| } | t|dz k(s'| dz|kDr/| dzj"j6zd k(r|j_t`stcjd dddg}| d#k(r| }ntgt| D]]})|d |)|z}*|d$|)|z}+ji| |)jkd |*j d%},|,jlj$jnz jlj$jpz},jljs|,d d }-jju|-| &d }-d }.g}/tK|*d%D]h\} }0|+| rV| d'k(r.|/jw|-jy|.d |.|0dz|0d fn#|/jw|-d |0d |.|.|0dzf|.|0dz }.j|jw|/`| d'k7r1t'jz|D12cgc]}1|1D]}2|2d c}2}1d (}j}| s|fSt|)S#1swYxYwcc}2}1w)*a: Function invoked when calling the VisualCloze pipeline for generation. Args: task_prompt (`str` or `List[str]`, *optional*): The prompt or prompts to define the task intention. content_prompt (`str` or `List[str]`, *optional*): The prompt or prompts to define the content or caption of the target image to be generated. image (`torch.Tensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.Tensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`): `Image`, numpy array or tensor representing an image batch to be used as the starting point. For both numpy array and pytorch tensor, the expected value range is between `[0, 1]` If it's a tensor or a list or tensors, the expected shape should be `(B, C, H, W)` or `(C, H, W)`. If it is a numpy array or a list of arrays, the expected shape should be `(B, H, W, C)` or `(H, W, C)`. num_inference_steps (`int`, *optional*, defaults to 50): The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. sigmas (`List[float]`, *optional*): Custom sigmas to use for the denoising process with schedulers which support a `sigmas` argument in their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed will be used. guidance_scale (`float`, *optional*, defaults to 30.0): 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. num_images_per_prompt (`int`, *optional*, defaults to 1): The number of images to generate per prompt. generator (`torch.Generator` or `List[torch.Generator]`, *optional*): One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation deterministic. latents (`torch.FloatTensor`, *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 will ge generated by sampling using the supplied random `generator`. prompt_embeds (`torch.FloatTensor`, *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. pooled_prompt_embeds (`torch.FloatTensor`, *optional*): Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, pooled text embeddings will be generated from `prompt` input argument. output_type (`str`, *optional*, defaults to `"pil"`): The output format of the generate image. Choose between [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`. return_dict (`bool`, *optional*, defaults to `True`): Whether or not to return a [`~pipelines.flux.FluxPipelineOutput`] instead of a plain tuple. joint_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`, *optional*): A function that calls at the end of each denoising steps during the inference. The function is called 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): Maximum sequence length to use with the `prompt`. Examples: Returns: [`~pipelines.flux.FluxPipelineOutput`] or `tuple`: [`~pipelines.flux.FluxPipelineOutput`] if `return_dict` is True, otherwise a `tuple`. When returning a tuple, the first element is a list with the generated images. )r&rxrrJFr2rvNr!scalerurw) rurvrwr&rxrKrIrJryc3K|]6}|D]/}|djzdz|djzdzz18yw)rrr!Nr )rrrrCs rErz9VisualClozeGenerationPipeline.__call__..3s\  !W-- - 2tAw$BWBW7W[\7\ ] ] s preprocessr2r]r^rr:r\rgetrrrr'r;nplinspacerrrhrrLrrrr-guidance_embedsrbfullfloat32expandra progress_barrrrgrstepbackendsmps is_availablelocalspopupdate XLA_AVAILABLExm mark_steprr unsqueezer(rrdecode postprocessrcrop concatenatemaybe_free_model_hooksr)3rCrvrwrrrrrIrr%r&rxr r rr rrJprocessor_outputrjrKryr image_seq_lenrrlatent_timesteprrnum_warmup_stepsrr+rtrlatent_model_input noise_pred latents_dtypecallback_kwargsrcallback_outputsbcur_image_sizecur_target_position cur_latent cur_imagercroppedr sub_imagearrs3` rE__call__z&VisualClozeGenerationPipeline.__call__sx    '!5/Q 3   .'=$//:: AVAV;  M * 6:FVWdFegj;kJ m , 8ZHXYfHgim=n-m<=J''?C>Y>Y>eD ' ' + +GT :ko 9=8J8J*?;(7+,<='!5"7 3!9K 9 5 +X *<8;    NN ! ! % %&:C @ NN ! ! % %&94 @ NN ! ! % %lC 8 NN ! ! % %k4 8  TZSaS!&9"9;NOgm); NN   * & &*.););:N:N \ * V $  . .     !22;O ; 7%'7s9~0CdnnFZFZ0ZZ\]^!)n    " " 2 2zz1#~fEMMZHw}}Q'78HH  %8  9- #\!),, #1>>88GMM!$4588G%*YY9M/NTU%V"!--"4%_%';*7$,+/+F+F %.   !( ..--j!WRW-XYZ[==M1~~))668")**]";'3&(O?9-3Xa[*9';D!Q'X$.229gFG$4$8$8-$XMI**A9I/IqSTuX\XfXfXlXlNlpqNq '')!LLNY, #- #f ( "E3w<( &!1,!?J!O&67H&I!j.&Y#!11'!*2F2Fq2I>[_[p[pqrtu (488??+I+IITXX__MiMii  HHOOJEOJ1M  00<>5!UTRSW_VZ[\V]:^+_`#NN9Qa[%%RVWXRY/BY5Y+Z[T!W$E % W%# &$e#5'^iT]'^SD '^ '^efg ##%8O!//e- #- #T(_s F(Z,5Z,.Z9 ,Z6)i)Nr!rGNN)r!N)Nr!NNrGN)NNNN)9__name__ __module__ __qualname____doc__model_cpu_offload_seq_optional_componentsrrrr r r r rrr7rr^rrrbrKrLrqrt FloatTensorfloatrTensorrrrr staticmethodrrrrrrrrrpropertyrrrrno_gradrEXAMPLE_DOC_STRINGboolrrrrN __classcell__)rDs@rEr$r$vs 8M(/:!'2!'!'$ !' ! !' ' !'%!',!'!'L)-%&#&)-'+ /c49n%/ #/! / & /  $ /j&')- *c49n%* #*& *d*.%&59<@#&&*X=S$s)^,X=3S >*X=c49n- X= & X= # X=  1 12X='u'8'89X=!X=UOX=v u||   8 !+/ >b@22(  ."#!"1If?BB$$,,##U]]_12.204-1#%(, $/0MQ/359<@%* ;?KO9B #&%K03S >*K0c49n-K0))* K0 ! K0 e% K0K0 (}K0E%//43H"HIJK0%++,K0  1 12K0'u'8'89K0c]K0K0!)c3h 8K0 'xc40@$0F'GH!K0"-1I#K0$!%K03K0rFr$)6typingrrrrrrnumpyr%rb transformersr r r r loadersrrrmodels.autoencodersrmodels.transformersr schedulersrutilsrrrrrrutils.torch_utilsrflux.pipeline_flux_fillrrrflux.pipeline_outputrpipeline_utilsr visualcloze_utilsr"torch_xla.core.xla_modelcore xla_modelr4r3 get_loggerrOrer[r$r}rFrEros>= VV\\099.[[5.3))MM   H %@FB 0 B 0rF