+h FSSKrSSKJrJrJrJrJrJr SSKr SSK r SSK r SSK JrJrJrJrJrJr SSKJrJr SSKJrJrJrJr SSKJrJr SSKJ r SS K!J"r"J#r#J$r$J%r%J&r&J'r' SS K(J)r) S S K*J+r+ S SK,J-r- \#"5(a SSK.J/s J0r1 Sr2OSr2\$Rf"\45r5Sr6/SQr7S#S\8S\8S\9S\94Sjjr:S$S\\8S\\\;\ Rx4S\\\8S\\\94Sjjr=S%S\ R|S\\ R~S\;4S jjr@"S!S"\+\\\\5rAg)&N)AnyCallableDictListOptionalUnion)CLIPImageProcessor CLIPTextModel CLIPTokenizerCLIPVisionModelWithProjectionT5EncoderModelT5TokenizerFast)PipelineImageInputVaeImageProcessor)FluxIPAdapterMixinFluxLoraLoaderMixinFromSingleFileMixinTextualInversionLoaderMixin) AutoencoderKLFluxTransformer2DModel)FlowMatchEulerDiscreteScheduler)USE_PEFT_BACKENDis_torch_xla_availableloggingreplace_example_docstringscale_lora_layersunscale_lora_layers) randn_tensor)DiffusionPipeline)FluxPipelineOutputTFa$ Examples: # Inpainting with text only ```py >>> import torch >>> from diffusers import FluxKontextInpaintPipeline >>> from diffusers.utils import load_image >>> prompt = "Change the yellow dinosaur to green one" >>> img_url = ( ... "https://github.com/ZenAI-Vietnam/Flux-Kontext-pipelines/blob/main/assets/dinosaur_input.jpeg?raw=true" ... ) >>> mask_url = ( ... "https://github.com/ZenAI-Vietnam/Flux-Kontext-pipelines/blob/main/assets/dinosaur_mask.png?raw=true" ... ) >>> source = load_image(img_url) >>> mask = load_image(mask_url) >>> pipe = FluxKontextInpaintPipeline.from_pretrained( ... "black-forest-labs/FLUX.1-Kontext-dev", torch_dtype=torch.bfloat16 ... ) >>> pipe.to("cuda") >>> image = pipe(prompt=prompt, image=source, mask_image=mask, strength=1.0).images[0] >>> image.save("kontext_inpainting_normal.png") ``` # Inpainting with image conditioning ```py >>> import torch >>> from diffusers import FluxKontextInpaintPipeline >>> from diffusers.utils import load_image >>> pipe = FluxKontextInpaintPipeline.from_pretrained( ... "black-forest-labs/FLUX.1-Kontext-dev", torch_dtype=torch.bfloat16 ... ) >>> pipe.to("cuda") >>> prompt = "Replace this ball" >>> img_url = "https://images.pexels.com/photos/39362/the-ball-stadion-football-the-pitch-39362.jpeg?auto=compress&cs=tinysrgb&dpr=1&w=500" >>> mask_url = ( ... "https://github.com/ZenAI-Vietnam/Flux-Kontext-pipelines/blob/main/assets/ball_mask.png?raw=true" ... ) >>> image_reference_url = ( ... "https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcTah3x6OL_ECMBaZ5ZlJJhNsyC-OSMLWAI-xw&s" ... ) >>> source = load_image(img_url) >>> mask = load_image(mask_url) >>> image_reference = load_image(image_reference_url) >>> mask = pipe.mask_processor.blur(mask, blur_factor=12) >>> image = pipe( ... prompt=prompt, image=source, mask_image=mask, image_reference=image_reference, strength=1.0 ... ).images[0] >>> image.save("kontext_inpainting_ref.png") ``` )) )))p) 0)@)p)P)r4)r3r2)r1r0)r/r.)r-r,)r+r*)r)r()r'r&)r%r$ base_seq_len max_seq_len base_shift max_shiftc4XC- X!- - nX5U-- nX-U-nU$N) image_seq_lenr5r6r7r8mbmus p/home/james-whalen/.local/lib/python3.13/site-packages/diffusers/pipelines/flux/pipeline_flux_kontext_inpaint.pycalculate_shiftrAs3  K$>?A%%A  Q B Inum_inference_stepsdevice timestepssigmasc UbUb [S5eUbS[[R"UR5R R 55;nU(d[SURS35eUR"S X2S.UD6 URn[U5nX14$UbS[[R"UR5R R 55;nU(d[SURS35eUR"S XBS.UD6 URn[U5nX14$UR"U4S U0UD6 URnX14$) a Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`. Args: scheduler (`SchedulerMixin`): The scheduler to get timesteps from. num_inference_steps (`int`): The number of diffusion steps used when generating samples with a pre-trained model. If used, `timesteps` must be `None`. device (`str` or `torch.device`, *optional*): The device to which the timesteps should be moved to. If `None`, the timesteps are not moved. timesteps (`List[int]`, *optional*): Custom timesteps used to override the timestep spacing strategy of the scheduler. If `timesteps` is passed, `num_inference_steps` and `sigmas` must be `None`. sigmas (`List[float]`, *optional*): Custom sigmas used to override the timestep spacing strategy of the scheduler. If `sigmas` is passed, `num_inference_steps` and `timesteps` must be `None`. Returns: `Tuple[torch.Tensor, int]`: A tuple where the first element is the timestep schedule from the scheduler and the second element is the number of inference steps. zYOnly one of `timesteps` or `sigmas` can be passed. Please choose one to set custom valuesrEzThe current scheduler class zx's `set_timesteps` does not support custom timestep schedules. Please check whether you are using the correct scheduler.)rErDrFzv's `set_timesteps` does not support custom sigmas schedules. Please check whether you are using the correct scheduler.)rFrDrDr;) ValueErrorsetinspect signature set_timesteps parameterskeys __class__rElen) schedulerrCrDrErFkwargsaccepts_timesteps accept_sigmass r@retrieve_timestepsrUs}>!3tuu'3w/@/@AXAX/Y/d/d/i/i/k+ll .y/B/B.CDab  M)MfM'' !)n  ))   C(9(9):Q:Q(R(](](b(b(d$ee .y/B/B.CD_`  GvGG'' !)n  ))  3MFMfM''  ))rBencoder_output generator sample_modec[US5(a!US:XaURRU5$[US5(a US:XaURR5$[US5(a UR$[ S5e)N latent_distsampleargmaxlatentsz3Could not access latents of provided encoder_output)hasattrrZr[moder]AttributeError)rVrWrXs r@retrieve_latentsras}~}--+2I))00;;  / /K84K))..00  + +%%%RSSrBcG^\rSrSrSrSrSS/rSS/rSWS \S \ S \ S \ S \ S\ S\S\S\4U4SjjjrSXS\\\\4S\S\S\\R0S\\R24 SjjrSYS\\\\4S\S\\R04SjjrSZS\\\\4S\\\\\4S\\R0S\S\\R8S\\R8S\S\\4SjjrSrSr S r!S[S!jr"\#S"5r$\#S#5r%\#S$5r&S%\RNS&\RP4S'jr)S(r*S)r+S*r,S+r-S\S%\\RNS,\S-\S.\S/\S0\S\R2S\R0S&\\\RP\\RP4S\\RNS1\\RN4S2jjr.S3r/\0S45r1\0S55r2\0S65r3\0S75r4\0S85r5\Rl"5\7"\85SSSSSSSS9SSS9SS:SS;SSSSSSSSSSSSS=4"S%\\9S1\\9S?\9S\\\\4S\\\\\4S@\\\\4SA\\\\\4SB\S/\\S0\\SC\SD\\SE\SF\\\SG\S\\S&\\\RP\\RP4S\\R8S\\R8S\\R8SH\\9SI\\\RNSJ\\9SK\\\RNSL\\R8SM\\R8SN\\SO\:SP\\;\\<4SQ\\=\\\;/S4SR\\S\SS\ST\:4DSUjj55r>SVr?U=r@$)]FluxKontextInpaintPipelinea The Flux Kontext pipeline for text-to-image generation. Reference: https://blackforestlabs.ai/announcing-black-forest-labs/ 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). z=text_encoder->text_encoder_2->image_encoder->transformer->vae image_encoderfeature_extractorr] prompt_embedsNrQvae text_encoder tokenizertext_encoder_2 tokenizer_2 transformerc ^>[T U]5 URUUUUUUUUU S9 [USS5(a/S[ UR R R5S- -OSUl[USS5(a UR R ROSUl [URS-S9Ul [URS-URSS S S 9Ul [US 5(a#URbURROS UlS Ulg)N) rhrirkrjrlrmrQrerfrhr r")vae_scale_factorFT)rqvae_latent_channels do_normalize do_binarizedo_convert_grayscalerjM)super__init__register_modulesgetattrrPrhconfigblock_out_channelsrqlatent_channelsrimage_processormask_processorr^rjmodel_max_lengthtokenizer_max_lengthdefault_sample_size) selfrQrhrirjrkrlrmrerfrOs r@ry#FluxKontextInpaintPipeline.__init__s(  %)##'/  W^^bdikoVpVpc$((//*L*L&MPQ&Q RvwCJ$PUW[B\B\txx>>bd0$BWBWZ[B[\/!22Q6 $ 4 4!%  07t[/I/IdnnNhDNN + +np !$' rBr"promptnum_images_per_promptmax_sequence_lengthrDdtypec U=(d URnU=(d URRn[U[5(aU/OUn[ U5n[U[ 5(aURXR5nURUSUSSSSS9nURnURUSSS9Rn U RSURS:ag[R"X5(dLURRU SS2URS - S245n [R!S US U 35 UR#UR%U5SS 9S n UR"RnU R%XTS9n U Rupn U R'S US 5n U R)Xb-U S5n U $)N max_lengthTFpt)paddingr truncation return_lengthreturn_overflowing_tokensreturn_tensorslongestrrr"zXThe following part of your input was truncated because `max_sequence_length` is set to tokens: output_hidden_statesrrrD)_execution_devicerir isinstancestrrPrmaybe_convert_promptrl input_idsshapetorchequal batch_decoderloggerwarningrktorepeatview)rrrrrDr batch_size text_inputstext_input_idsuntruncated_ids removed_textrg_seq_lens r@_get_t5_prompt_embeds0FluxKontextInpaintPipeline._get_t5_prompt_embeds(s14110**00'44&&[ d7 8 8..v7G7GHF&&  *&+' %..**69UY*Zdd   $(<(>66q$JcJcfgJgjlJlGl7mnL NN--.i ~G )).*;*;F*CZ_)` &33 %((t/@/@/F/Fv(V &,,Q0EF %**:+MrR rBprompt_2pooled_prompt_embeds lora_scalec U=(d URnUbw[U[5(abXlURb![ (a[ URU5 URb![ (a[ URU5 [U[5(aU/OUnUcHU=(d Un[U[5(aU/OUnURUUUS9nURUUUUS9nURb6[U[5(a![ (a[URU5 URb6[U[5(a![ (a[URU5 URbURROURRn [R"UR SS5R#X9S9n XVU 4$)a? Args: prompt (`str` or `List[str]`, *optional*): prompt to be encoded prompt_2 (`str` or `List[str]`, *optional*): The prompt or prompts to be sent to the `tokenizer_2` and `text_encoder_2`. If not defined, `prompt` is used in all text-encoders device: (`torch.device`): torch device num_images_per_prompt (`int`): number of images that should be generated per prompt prompt_embeds (`torch.FloatTensor`, *optional*): Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, text embeddings will be generated from `prompt` input argument. pooled_prompt_embeds (`torch.FloatTensor`, *optional*): Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, pooled text embeddings will be generated from `prompt` input argument. lora_scale (`float`, *optional*): A lora scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded. )rrDr)rrrrDr"rrDr)rrr _lora_scalerirrrkrrrrrrmrzerosrr) rrrrDrrgrrrrtext_idss r@ encode_prompt(FluxKontextInpaintPipeline.encode_prompts@1411  !j7J&K&K)   ,1A1A!$"3"3Z@"".3C3C!$"5"5zB'44&&  )6H%/#%>%>zHH$(#?#?&;$@$ !66&;$7 7M    ($ 3449I9I#D$5$5zB    *$ 3449I9I#D$7$7D+/+<+<+H!!''dN^N^NdNd;;}2215q9< U$) NzK`ip_adapter_image` must have same length as the number of IP Adapters. Got z images and z IP Adapters.r"zR`ip_adapter_image_embeds` must have same length as the number of IP Adapters. Got z image embeds and rr)rD) rlistrPrmencoder_hid_projnum_ip_adaptersrHrappendrcatr)rip_adapter_imageip_adapter_image_embedsrDrrsingle_ip_adapter_imagesingle_image_embedss r@prepare_ip_adapter_image_embeds:FluxKontextInpaintPipeline.prepare_ip_adapter_image_embedss  " *.55$4#5 #$(8(8(I(I(Y(YY abefvbwaxyEFJFVFVFgFgFwFwExxEF,<'&*&7&78OYZ&[###$7a$@A,<5t<<+B*C'*+t/?/?/P/P/`/`` hilnEjFiGGYZ^ZjZjZ{Z{ZKZKYLLYZ(?###$78(?#%#/ "'))-@,ADY,Y_`"a "5"8"8"8"G  # * *+> ?$0 '&rBcN[X-U5n[[X- S55nURRXPRR -Sn[ URS5(a1URRXPRR -5 XaU- 4$)Nrset_begin_index)minintmaxrQrEorderr^r)rrCstrengthrD init_timestept_startrEs r@ get_timesteps(FluxKontextInpaintPipeline.get_timestepss/:>#4 5 5 NN * *7^^5I5I+I J777rBc ^US:dUS:a[SU35eUTRS--S:wdUTRS--S:wa,[RSTRS-SUSUS35 UbX[ U4S jU55(d>[S TR S UVs/sHnUTR ;dMUPM sn35eUbU b[S US U S35eUbU b[SUS U S35eUcU c [S5eUbA[ U[5(d,[ U[5(d[S[U535eUbA[ U[5(d,[ U[5(d[S[U535eU bU b[SU SU S35eU bU b[SU SU S35eU bCU b@U RU R:wa&[SU RSU RS35eU bU c [S5eU bUc [S5eUb[ U[RR5(d[S[U5S35e[ U[RR5(d[S[U5S35eUS:wa[SUS35eUbUS:a[S U35eggs snf)!Nrr"z2The value of strength should in [0.0, 1.0] but is r z-`height` and `width` have to be divisible by z but are  and z(. Dimensions will be resized accordinglyc3@># UHoTR;v M g7fr:)_callback_tensor_inputs).0krs r@ :FluxKontextInpaintPipeline.check_inputs..0sF 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 `prompt_2`: 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 z4`prompt_2` has to be of type `str` or `list` but is z'Cannot forward both `negative_prompt`: z and `negative_prompt_embeds`: z)Cannot forward both `negative_prompt_2`: zu`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but got: `prompt_embeds` z != `negative_prompt_embeds` .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`.zIf `negative_prompt_embeds` are provided, `negative_pooled_prompt_embeds` also have to be passed. Make sure to generate `negative_pooled_prompt_embeds` from the same text encoder that was used to generate `negative_prompt_embeds`.zJThe image should be a PIL image when inpainting mask crop, but is of type zOThe mask image should be a PIL image when inpainting mask crop, but is of type pilz@The output type should be PIL when inpainting mask crop, but is rz8`max_sequence_length` cannot be greater than 512 but is ) rHrqrrallrrrrtyperPILImage)rrrr mask_imagerheightwidth output_typenegative_promptnegative_prompt_2rgnegative_prompt_embedsrnegative_pooled_prompt_embeds"callback_on_step_end_tensor_inputspadding_mask_croprrs` r@ check_inputs'FluxKontextInpaintPipeline.check_inputss ( a<8a<QRZQ[\] ] T**Q. /1 4AVAVYZAZ8[_`8` NN?@U@UXY@Y?ZZcdjckkpqvpwx`a  . 9#F 7YF C C DTEaEaDbbn|^pH|^vwbckokGkGbGpq|^pHoIJ   -";08N}o^00  !m&?28*RSTVZSZ>[#[ #3F#;ell5>QRVXYRY>Z#Z RbRhRhO7O+33 " :>rBcURXUS-SUS-S5nURSSSSSS5nURXS-US--US-5nU$)Nr rr"r)rpermuter)r]rnum_channels_latentsrrs r@ _pack_latents(FluxKontextInpaintPipeline._pack_latentssg,,z1aQVZ[Q[]^_//!Q1a3//*{uz.JL`cdLderBcURupEnS[U5US---nS[U5US---nURXAS-US-US-SS5nURSSSSSS5nUR XFS-X5nU$)Nr rrrr"r)rrrrr)r]rrrqr num_patcheschannelss r@_unpack_latents*FluxKontextInpaintPipeline._unpack_latentss-4MM) c&k&6&:;<SZ$4q$89:,,zQ; HPQMSTVWX//!Q1a3//*5.A6QrBrrWc [U[5(ag[URS5Vs/sH.n[ UR R XUS-5X#SS9PM0 nn[R"USS9nO$[ UR R U5USS9nX@R RR- UR RR-nU$s snf)Nrr"r\)rWrXr) rrrangerrarhencoderrr| shift_factorscaling_factor)rrrWi image_latentss r@_encode_vae_image,FluxKontextInpaintPipeline._encode_vae_images i & &u{{1~..A!1q51A!Bilhpq. "IIm;M,TXX__U-CyfnoM&)E)EEIgIgg s5C&c8URR5 g)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)rhenable_slicingrs r@enable_vae_slicing-FluxKontextInpaintPipeline.enable_vae_slicing !rBc8URR5 g)z Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to computing decoding in one step. N)rhdisable_slicingrs r@disable_vae_slicing.FluxKontextInpaintPipeline.disable_vae_slicings   "rBc8URR5 g)z 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)rh enable_tilingrs r@enable_vae_tiling,FluxKontextInpaintPipeline.enable_vae_tilings  rBc8URR5 g)z Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to computing decoding in one step. N)rhdisable_tilingrs r@disable_vae_tiling-FluxKontextInpaintPipeline.disable_vae_tilingrrBtimesteprrrrimage_referencec [U [5(a*[U 5U:wa[S[U 5SUS35eS[ U5UR S---nS[ U5UR S---nX4XV4n S=pUbUR XS9nURSUR:waURXS9n OUn X=RS:a@X=RS-S:Xa+X=RS-n[R"U /U-SS 9n O\X=RS:a4X=RS-S:wa[S U RSS US 35e[R"U /SS 9n S=nnU bU R XS9n U RSUR:waURXS9nOU nUURS:aBUURS-S:Xa,UURS-n[R"U/U-SS 9nO^UURS:a5UURS-S:wa[S URSS US 35e[R"U/SS 9nURX5S-US-X5nU c'[XXS9nURRXU5n OU R XS9nUn U RSSunnUR!XUUU5n URUUS-US-X5nSUS'UbFURSSunnUR!UUUUU5nURUUS-US-X5nSUS'UR!UX4XV5nUR!XXEU5n XUUUUU4$)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.r rr")rrWrrz'Cannot duplicate `image` of batch size z to z text prompts.z1Cannot duplicate `image_reference` of batch size )rWrDr).r)rrrPrHrrqrrr~rrrrrrQ scale_noiser )rrr+rrrrrrDrWr]r,rr image_idsadditional_image_per_promptimage_reference_latentsimage_reference_ids latent_idsnoiseimage_latent_heightimage_latent_widthimage_reference_latent_heightimage_reference_latent_widths r@prepare_latents*FluxKontextInpaintPipeline.prepare_latentssU i & &3y>Z+GA#i.AQR&<'gi c&kd&;&;a&?@ASZD$9$9A$=>?6A%)(  HHFH8E{{1~!5!55 $ 6 6U 6 X % //22zDWDWXYDZ7Z^_7_.8Q>QRS>T=UUYZdYeest!& =/q A 9=<"5  &-000LO$$Q'4+?+??*.*@*@*@*l'*9'399!<<NeNkNklmNnAnrsAs.8>:PgPmPmnoPpCptuCu GH_HeHefgHhGiimnxmyzHI+0))5L4MST*U'33J! UVWZY_g ? FXEnn00%PGJJfJ:EG2?2E2Eab2I//** ';=PRd 22 +q02D2I6  & " .JaJgJghihjJk G )+G&*&8&8'$-, ' ##'"@"@9Q>@\`a@aci# +,  '""5*FZ$$W:NX]^'> IWjlqqqrBc LS[U5URS---nS[U5URS---n[RRR XU4S9nUR XS9nX5-nUR XS9nURSS:XaUn Of[URRU5U S9n XRRR- URRR-n URSU:aWX1RS-S:Xd[SUS URSS 35eURX1RS-SSS5nU RSU:aWX;RS-S:Xd[S US U RSS 35eU RX;RS-SSS5n U R XS9n UR!U UUUU5n UR!URSUSS5UUUU5nX4$) Nr )sizerr"rp)rWrzvThe passed mask and the required batch size don't match. Masks are supposed to be duplicated to a total batch size of z, but zo masks were passed. Make sure the number of masks that you pass is divisible by the total requested batch size.zyThe passed images and the required batch size don't match. Images are supposed to be duplicated to a total batch size of zq images were passed. Make sure the number of images that you pass is divisible by the total requested batch size.)rrqrnn functional interpolaterrrarhrr|rrrHrr ) rmask masked_imagerrrrrrrDrWmasked_image_latentss r@prepare_mask_latents/FluxKontextInpaintPipeline.prepare_mask_latents6sdc&kd&;&;a&?@ASZD$9$9A$=>?xx""..t5/.Jwwfw27 #fB   a B &#/ #3DHHOOL4Q]f#g %xx'C'CC..$/ ::a=: % 1 -2 ..8\ 1 O^^ ;;zZZ]:Aq!DD  % %a (: 5 : :1 ==B 11; FCWC]C]^_C`Babtt $8#>#>zMgMghiMj?jlmoprs#t  466f6R#11     !! KK/A 6     ))rBcUR$r:)_guidance_scalers r@guidance_scale)FluxKontextInpaintPipeline.guidance_scales###rBcUR$r:)_joint_attention_kwargsrs r@joint_attention_kwargs1FluxKontextInpaintPipeline.joint_attention_kwargss+++rBcUR$r:)_num_timestepsrs r@ num_timesteps(FluxKontextInpaintPipeline.num_timestepss"""rBcUR$r:)_current_timesteprs r@current_timestep+FluxKontextInpaintPipeline.current_timesteps%%%rBcUR$r:) _interruptrs r@ interrupt$FluxKontextInpaintPipeline.interrupts rB?g @rTirrrtrue_cfg_scalerrrCrFrGrrnegative_ip_adapter_image negative_ip_adapter_image_embedsrrr return_dictrKcallback_on_step_endrmax_area _auto_resizec#^WU =(d URUR-n U =(d URUR-n Xn$n#X- mW[U!TW-S-5n [U!TW- S-5n URS-n%U U%-U%-n U U%-U%-n U U#:wdU U$:wa[R SU SU S35 UR UUUUU U U UUUUUUUUU U S9 XlUUlSUlSUl UGbt[U[R5(a URS 5UR:XGd5[U[5(aJ[US [R5(a(US R S :Xa[R""US S 9n[U[5(aUS OUn&UR$R'U&5un'n(U(U'- mWU"(a[)UW4S j[*55un)n(n'U(U%-U%-n(U'U%-U%-n'UR$R-UU'U(5nU(n U'n U bUR.R1X:XS9n*Sn+OSn*Sn+UR$R3UU'U(U*U+S9nO [5S5eUR7[R8S9n,UGbL[U[R5(a URS 5UR:XGd [U[5(aJ[US [R5(a(US R S :Xa[R""US S 9n[U[5(aUS OUn-UR$R'U-5un.n/U/U.- mWU"(a[)UW4Sj[*55un)n/n.U/U%-U%-n/U.U%-U%-n.UR$R-UU.U/5nUR$R3UU.U/U*U+S9nOSnUb[U[:5(aS n0O3Ub![U[5(a [=U5n0OUR>S n0UR@n1URBbURBRESS5OSn2USL=(d USL=(a USLn3US :=(a U3n4URGUUUUU1UU U2S9unnn5U4(aURGUUUUU1UU U2S9unnn6UcRJ"SS U - U 5OUn[MU 5UR-S-[MU 5UR-S--n7[OU7URPRRRESS5URPRRRESS5URPRRRESS5URPRRRESS55n8[UURPU U1UU8S9un9n URWXU15un9n U S :a[5S U S!U S"35eU9SS RYU0U-5n:[[[=U95XRPR\-- S 5n;[=U95Ul/UR`RRRbS -nn?n@nAnBUAb[R""U?WA/S S 9n?OW@b[R""U?W@/S S 9n?UR.R3X9U U+U*S#9nCU,UCS:-nDURiUCUDU0US 5nFOSnFUcUbPUcMUcJ[HRp"XS%4[HRrS9nU/UR`RtRv-nOUUcRUcOUcUbI[HRp"XS%4[HRrS9nU/UR`RtRv-nURBc0UlSnGSnHUcUbURyUUU1U0U-5nGUcUbURyUUU1U0U-5nHUR{U S&9nI[}U95GHunJnKUR~(aMWKUlWGbWGURS''UnLU>b[R""UU>/S S 9nLOU=b[R""UU=/S S 9nLWKRoUR>S 5R7URf5nMURaWLUMS(- WFUUU5U?URBSS)9 S nNUNSS2SURS 524nNU4(a^WHbWHURS''URaWLWMS(- WFUUW6U?URBSS)9 S nOUOSS2SURS 524nOUOUWNUO- --nNURfnPURPRWNWKUSS*9S nU=nQWEnRWJ[=U95S - :a:U9WJS -nSURPRWQ[R"US/5WB5nQS WR- WQ-URU--nURfWP:wa>[RRR5(aUR7WP5nUbJ0nTUHnU[5UUWTUU'M U"UWJWKWT5nVUVRS+U5nUVRS,U5nWJ[=U95S - :Xd)WJS -U;:a0WJS -URPR\-S :XaWIR5 [(dGM[R"5 GM SSS5 SUlUS-:XaUnOURUXUR5nUURRRR- URRRR-nURRUSS*9S nUR$RUUS.9nUR5 U(dU4$[US/9$!,(df  N=f)0a) Function invoked when calling the pipeline for generation. Args: 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 be inpainted (which parts of the image to be masked out with `mask_image` and repainted according to `prompt` and `image_reference`). 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)` It can also accept image latents as `image`, but if passing latents directly it is not encoded again. image_reference (`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 the masked area. 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)` It can also accept image latents as `image`, but if passing latents directly it is not encoded again. mask_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 mask `image`. White pixels in the mask are repainted while black pixels are preserved. If `mask_image` is a PIL image, it is converted to a single channel (luminance) before use. If it's a numpy array or pytorch tensor, it should contain one color channel (L) instead of 3, so the expected shape for pytorch tensor would be `(B, 1, H, W)`, `(B, H, W)`, `(1, H, W)`, `(H, W)`. And for numpy array would be for `(B, H, W, 1)`, `(B, H, W)`, `(H, W, 1)`, or `(H, W)`. prompt (`str` or `List[str]`, *optional*): The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`. instead. prompt_2 (`str` or `List[str]`, *optional*): The prompt or prompts to be sent to `tokenizer_2` and `text_encoder_2`. If not defined, `prompt` is will be used instead. 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 `true_cfg_scale` is not greater than `1`). negative_prompt_2 (`str` or `List[str]`, *optional*): The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and `text_encoder_2`. If not defined, `negative_prompt` is used in all the text-encoders. true_cfg_scale (`float`, *optional*, defaults to 1.0): True classifier-free guidance (guidance scale) is enabled when `true_cfg_scale` > 1 and `negative_prompt` is provided. height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor): The height in pixels of the generated image. This is set to 1024 by default for the best results. width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor): The width in pixels of the generated image. This is set to 1024 by default for the best results. strength (`float`, *optional*, defaults to 1.0): Indicates extent to transform the reference `image`. Must be between 0 and 1. `image` is used as a starting point and more noise is added the higher the `strength`. The number of denoising steps depends on the amount of noise initially added. When `strength` is 1, added noise is maximum and the denoising process runs for the full number of iterations specified in `num_inference_steps`. A value of 1 essentially ignores `image`. padding_mask_crop (`int`, *optional*, defaults to `None`): The size of margin in the crop to be applied to the image and masking. If `None`, no crop is applied to image and mask_image. If `padding_mask_crop` is not `None`, it will first find a rectangular region with the same aspect ration of the image and contains all masked area, and then expand that area based on `padding_mask_crop`. The image and mask_image will then be cropped based on the expanded area before resizing to the original image size for inpainting. This is useful when the masked area is small while the image is large and contain information irrelevant for inpainting, such as background. 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 3.5): Embedded guidance scale is enabled by setting `guidance_scale` > 1. Higher `guidance_scale` encourages a model to generate images more aligned with `prompt` at the expense of lower image quality. Guidance-distilled models approximates true classifier-free guidance for `guidance_scale` > 1. Refer to the [paper](https://huggingface.co/papers/2210.03142) to learn more. 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. ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters. ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*): Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of IP-adapters. Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. If not provided, embeddings are computed from the `ip_adapter_image` input argument. negative_ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters. negative_ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*): Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of IP-adapters. Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. If not provided, embeddings are computed from the `ip_adapter_image` input argument. negative_prompt_embeds (`torch.FloatTensor`, *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. negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*): Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_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`. max_area (`int`, defaults to `1024 ** 2`): The maximum area of the generated image in pixels. The height and width will be adjusted to fit this area while maintaining the aspect ratio. 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 z6Generation `height` and `width` have been adjusted to rz to fit the model requirements.) rrrrgrrrrrrNFr"rrrc3L># UHup[TX- - 5X4v M g7fr:absrwh aspect_ratios r@r6FluxKontextInpaintPipeline.__call__..ws(3B_$!S-.5B_!$)padfilldefault) crops_coords resize_modez/image must be provided correctly for inpainting)rc3L># UHup[TX- - 5X4v M g7fr:rergs r@rrks)GB_$!S-.5B_rlscale)rrrgrrDrrrrYbase_image_seq_lenmax_image_seq_lenr7r8ffffff?)rFr?z?After adjusting the num_inference_steps by strength parameter: z!, the number of pipelinesteps is z4 which is < 1 and not appropriate for this pipeline.)rrrqrprr)totalri) hidden_statesr+guidancepooled_projectionsencoder_hidden_statestxt_idsimg_idsrKr^)r^r]rglatent)r)images)TrrqroundrrrrFrJrRrVrrrr<r~rndimrrget_default_height_widthrPREFERRED_KONTEXT_RESOLUTIONSresizerget_crop_region preprocessrHrfloat32rrPrrrKgetrnplinspacerrArQr|rUrrrrrNrm in_channelsr9rrCguidance_embedsfullexpandruint8rrr progress_bar enumeraterWstepr.tensorbackendsmps is_availablelocalspopupdate XLA_AVAILABLExm mark_steprrhrrdecode postprocessmaybe_free_model_hooksr#)Xrrr,rrrrrr[rrrrrCrFrGrrWr]rgrrrr\r]rrrr^rKr_rrr`raoriginal_heightoriginal_width multiple_ofimg image_height image_widthrrprq init_image img_referenceimage_reference_heightimage_reference_widthrrDrhas_neg_prompt do_true_cfgrnegative_text_idsr<r?rElatent_timestepnum_warmup_stepsrrr1r3r/r2r4mask_conditionrAr@r{rnegative_image_embedsrrtlatent_model_inputr+ noise_predneg_noise_pred latents_dtypeinit_latents_proper init_masknoise_timestepcallback_kwargsrcallback_outputsrjsX @r@__call__#FluxKontextInpaintPipeline.__call__sf `K433d6K6KKI11D4I4II*0~ x,.367</C78++a/ ${2;&4 _ $(? NNHPUV[U\\{|         #+/'#9!5*G/Q/ 3#  ( .'=$!%  j &E&E%**UV-[_[o[oJo%&&:eAh +M+MRWXYRZR_R_cdRd %Q/(55%(5C(,(<(<(U(UVY(Z %L+&5L/23B_30,; &4{BK';6DL((//|[QE E!F!,#22BB:V\Bt $ # ' ((33|[|Yd4ENO OXXEMMX2   &  5 5/:N:Nq:QUYUiUi:i?D11q15<<@@#A&++q0"'))O"C2<_d2S2SOA.YhM<@O#O  *VS"9"9J  Jvt$<$<VJ&,,Q/J''?C>Y>Y>eD ' ' + +GT :ko )4 "$ . \3PX\3\ %q(;^   '!5"7 3!      ""&*4%B&;$7%#   &-!TZSaS!&9"9;NOgmV(=(==Bs5zUYUjUjGjnoGop   NN ! ! % %&:C @ NN ! ! % %&94 @ NN ! ! % %lC 8 NN ! ! % %k4 8  *< NN   * & &*.););>)*&+NZD001JK%,"*6).G=T3U[\)]&".).G]3KQR)S&88GMM!$4588G!--"4%_%';*7$&+/+F+F %.   (+:J-KU+'y=,??)gBUU==M1~~))6688")**]";'3&(O?-3Xa[*@';D!Q'X$.229gFG$4$8$8-$XMI**A9I/IqSTuX\XfXfXlXlNlpqNq '') =LLN_-:d"& ( "E**7F4CXCXYG!?!??488??C_C__GHHOOGO?BE((44U 4TE ##%8O!//E: 9s'K q$7q$$ q2) rRrFrVrJrrNrrr~rrrq)NN)Nr"rNN)r"N)NNr"NNrN) NNNNNNNNN)NNN)A__name__ __module__ __qualname____firstlineno____doc__model_cpu_offload_seq_optional_componentsrrrr r r rrr r ryrrrrrrrDrrr FloatTensorfloatrrrrr staticmethodrr rr Generatorrrr!r%r)r9rCpropertyrGrKrOrSrWno_gradrEXAMPLE_DOC_STRINGrboolrrrr__static_attributes__ __classcell__)rOs@r@rcrcs4\+-@A(/:8<04*'2*'*'$ *' ! *' ' *'%*',*'5*'.*'*'^)-%&#&)-'+ /c49n%/ #/! / & /  $ /j&')- *c49n%* #*& *`59)-%&59<@#&&*M=c49n%M=5d3i01M=& M= # M=   1 12 M='u'8'89M=!M=UOM=` "'J 8,#!&*+/ %`oD ? ?   u||  "#!""NR*.26er %erer er " er  erer{{er erE%//43H"HIJer%,,'er"%,,/erPG*R$$,,##&& ]]_12/38<)-(,4815=A # $#+/#%(, #/0MQ/359<@9=@DBFIM>BEI%* ;?KO9B #&!G_0*+_0""45_0' _0 c49n% _0 5d3i01 _0sDI~._0$E#tCy.$9:_0_0 _0}_0_0$C=_0!_0e%_0 !_0" (}#_0$E%//43H"HIJ%_0&%++,'_0(  1 12)_0*'u'8'89+_0,##56-_0."*$u||*#?1_02+34 3E*F3_04!)):): ;5_06(00A0A'B7_08c]9_0:;_0<!)c3h 8=_0>'xc40@$0F'GH?_0@-1IA_0B!C_0DE_0FG_03_0rBrc)rurwrcrx)NNNN)Nr[)BrJtypingrrrrrrnumpyr PIL.Imagerr transformersr r r r r rrrrloadersrrrrmodelsrr schedulersrutilsrrrrrrutils.torch_utilsrpipeline_utilsr!pipeline_outputr#torch_xla.core.xla_modelcore xla_modelrr get_loggerrrrrrrrArrDrUrrrarcr;rBr@rst== Epp;9../))MM   H %:x!.       *.15%)$( 8*!#8* U3 ,- .8*S " 8* T%[ ! 8*zck TLL T-5eoo-F T\_ T^0 ^0rB