bi rddlmZddlmZddlmZddlmZddlm Z ddl m Z dd l m Z dd lmZdd lmZmZdd lmZmZmZmZmZmZmZmZmZdd lmZmZddl m!Z!m"Z"m#Z#ddl$m%Z%m&Z&m'Z'm(Z(m)Z)m*Z*m+Z+m,Z,m-Z-m.Z.ddl/m0Z0ddl1m2Z2m3Z3m4Z4m5Z5m6Z6m7Z7ddl8m9Z9m:Z:m;Z;mZ>ddl?m@Z@mAZAddlBmCZCmDZDddlEmFZFddlGmHZHddlImJZJmKZKmLZLmMZMmNZNmOZOmPZPmQZQmRZRmSZSmTZTmUZUmVZVmWZWmXZXddlYmZZZm[Z[ddl\m]Z]ddl^m_Z_m`Z`ddlambZbmcZcmdZdddlemfZfmgZgmhZhddlimjZjmkZkmlZlddlmmnZnmoZoegdedfd elfd!ehfd"eNfd#e#fd$e0fd%eJfd&e2fd'e9fd(eAfd)efd*efd+efd,efd-enfd.e_fd/eDfd0eZfd1e[fd2e]fd3eLfd4eSfd5ePfd6eXfd7eUfd8eKfd9e fd:e.fd;e*fdeFfd?eHfd@e fdAefdBefdCefZpedebfd ejfd!effd"eMfd#e!fd&e3fd'e:fd(e@fd)efd4eQfd*efd+efd6eVfd7eTfd/eCfd:e+fdfd-eofd.e`fgZsed&e4fd'e;fgZted&e6fd'e=fgZue rddDlvmwZwmxZxddElImyZyexepdF<eyepdG<eweqdF<epeqereseteugZzdHZ{dIZ|dSdJe}fdKZ~GdLdMeZGdNdOeZGdPdQeZyR)T) OrderedDict)validate_hf_hub_args) ConfigMixin)ControlNetUnionModel)is_sentencepiece_available)AuraFlowPipeline)ChromaPipeline)CogView3PlusPipeline)CogView4ControlPipelineCogView4Pipeline) (StableDiffusionControlNetImg2ImgPipeline(StableDiffusionControlNetInpaintPipeline!StableDiffusionControlNetPipeline*StableDiffusionXLControlNetImg2ImgPipeline*StableDiffusionXLControlNetInpaintPipeline#StableDiffusionXLControlNetPipeline/StableDiffusionXLControlNetUnionImg2ImgPipeline/StableDiffusionXLControlNetUnionInpaintPipeline(StableDiffusionXLControlNetUnionPipeline),StableDiffusion3ControlNetInpaintingPipeline"StableDiffusion3ControlNetPipeline)IFImg2ImgPipelineIFInpaintingPipeline IFPipeline) FluxControlImg2ImgPipelineFluxControlInpaintPipelineFluxControlNetImg2ImgPipelineFluxControlNetInpaintPipelineFluxControlNetPipelineFluxControlPipelineFluxImg2ImgPipelineFluxInpaintPipelineFluxKontextPipeline FluxPipeline)HunyuanDiTPipeline)KandinskyCombinedPipeline KandinskyImg2ImgCombinedPipelineKandinskyImg2ImgPipeline KandinskyInpaintCombinedPipelineKandinskyInpaintPipelineKandinskyPipeline)KandinskyV22CombinedPipeline#KandinskyV22Img2ImgCombinedPipelineKandinskyV22Img2ImgPipeline#KandinskyV22InpaintCombinedPipelineKandinskyV22InpaintPipelineKandinskyV22Pipeline)Kandinsky3Img2ImgPipelineKandinsky3Pipeline)%LatentConsistencyModelImg2ImgPipelineLatentConsistencyModelPipeline)LuminaPipeline)Lumina2Pipeline)HunyuanDiTPAGPipelinePixArtSigmaPAGPipelineSanaPAGPipeline"StableDiffusion3PAGImg2ImgPipelineStableDiffusion3PAGPipeline+StableDiffusionControlNetPAGInpaintPipeline$StableDiffusionControlNetPAGPipeline!StableDiffusionPAGImg2ImgPipeline!StableDiffusionPAGInpaintPipelineStableDiffusionPAGPipeline-StableDiffusionXLControlNetPAGImg2ImgPipeline&StableDiffusionXLControlNetPAGPipeline#StableDiffusionXLPAGImg2ImgPipeline#StableDiffusionXLPAGInpaintPipelineStableDiffusionXLPAGPipeline)PixArtAlphaPipelinePixArtSigmaPipeline) SanaPipeline)StableCascadeCombinedPipelineStableCascadeDecoderPipeline)StableDiffusionImg2ImgPipelineStableDiffusionInpaintPipelineStableDiffusionPipeline)StableDiffusion3Img2ImgPipelineStableDiffusion3InpaintPipelineStableDiffusion3Pipeline) StableDiffusionXLImg2ImgPipeline StableDiffusionXLInpaintPipelineStableDiffusionXLPipeline)WuerstchenCombinedPipelineWuerstchenDecoderPipelinezstable-diffusionzstable-diffusion-xlzstable-diffusion-3zstable-diffusion-3-pagifhunyuanz hunyuan-pag kandinsky kandinsky22 kandinsky3zstable-diffusion-controlnetzstable-diffusion-xl-controlnetz$stable-diffusion-xl-controlnet-unionzstable-diffusion-3-controlnet wuerstchencascadelcmz pixart-alphaz pixart-sigmasanazsana-pagzstable-diffusion-pagzstable-diffusion-controlnet-pagzstable-diffusion-xl-pagz"stable-diffusion-xl-controlnet-pagzpixart-sigma-pagauraflowfluxz flux-controlzflux-controlnetz flux-kontextluminalumina2chromacogview3cogview4zcogview4-control)KolorsImg2ImgPipelineKolorsPipeline)KolorsPAGPipelinekolorsz kolors-pagc0|tjvrtt|jdS|t jvrtt |jdS|tjvrtt|jdSy)NF)throw_error_if_not_exist) *_AUTO_TEXT2IMAGE_DECODER_PIPELINES_MAPPINGvalues_get_task_class!AUTO_TEXT2IMAGE_PIPELINES_MAPPING__name__+_AUTO_IMAGE2IMAGE_DECODER_PIPELINES_MAPPING"AUTO_IMAGE2IMAGE_PIPELINES_MAPPING'_AUTO_INPAINT_DECODER_PIPELINES_MAPPINGAUTO_INPAINT_PIPELINES_MAPPING) pipeline_clss \/home/cdr/jupyterlab/.venv/lib/python3.12/site-packages/diffusers/pipelines/auto_pipeline.py_get_connected_pipelinerzsAHHJJ -|/D/D_d  BIIKK . 0E0E`e  >EEGG=|?T?TotuuHcttD]/}|jD]\}}|j|k(s|ccS1y)N)SUPPORTED_TASKS_MAPPINGSitemsrs)pipeline_class_name task_mapping model_namepipelines ry _get_modelrsC0" $0$6$6$8 " J  $77!! ""r{rncpt|}||j|d}||S|rtd|d|y)Nz-AutoPipeline can't find a pipeline linked to z for )rget ValueError)mappingrrnr task_classs ryrqrqsU/0J[[T2  ! HI\H]]bcmbnopp r{cDeZdZdZdZdZeedZedZ y)AutoPipelineForText2Imagea6 [`AutoPipelineForText2Image`] is a generic pipeline class that instantiates a text-to-image pipeline class. The specific underlying pipeline class is automatically selected from either the [`~AutoPipelineForText2Image.from_pretrained`] or [`~AutoPipelineForText2Image.from_pipe`] methods. This class cannot be instantiated using `__init__()` (throws an error). Class attributes: - **config_name** (`str`) -- The configuration filename that stores the class and module names of all the diffusion pipeline's components. model_index.jsonct|jjd|jjd|jjdNz+ is designed to be instantiated using the `z5.from_pretrained(pretrained_model_name_or_path)` or `z.from_pipe(pipeline)` methods.EnvironmentError __class__rsselfargskwargss ry__init__z"AutoPipelineForText2Image.__init__#R~~&&'(..1123''((F H  r{c V|jdd}|jdd}|jdd}|jdd}|jdd}|jdd}||||||d } |j|fi| } | d } d | vrd } nd } d |vr>t|d tr| d j | d} n| d j | d} d|vr%|jd} | r| j | d} t t | }i| |}|j|fi|S)u( Instantiates a text-to-image Pytorch diffusion pipeline from pretrained pipeline weight. The from_pretrained() method takes care of returning the correct pipeline class instance by: 1. Detect the pipeline class of the pretrained_model_or_path based on the _class_name property of its config object 2. Find the text-to-image pipeline linked to the pipeline class using pattern matching on pipeline class name. If a `controlnet` argument is passed, it will instantiate a [`StableDiffusionControlNetPipeline`] object. The pipeline is set in evaluation mode (`model.eval()`) by default. If you get the error message below, you need to finetune the weights for your downstream task: ``` Some weights of UNet2DConditionModel were not initialized from the model checkpoint at stable-diffusion-v1-5/stable-diffusion-v1-5 and are newly initialized because the shapes did not match: - conv_in.weight: found shape torch.Size([320, 4, 3, 3]) in the checkpoint and torch.Size([320, 9, 3, 3]) in the model instantiated You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. ``` Parameters: pretrained_model_or_path (`str` or `os.PathLike`, *optional*): Can be either: - A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline hosted on the Hub. - A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights saved using [`~DiffusionPipeline.save_pretrained`]. torch_dtype (`torch.dtype`, *optional*): Override the default `torch.dtype` and load the model with another dtype. force_download (`bool`, *optional*, defaults to `False`): Whether or not to force the (re-)download of the model weights and configuration files, overriding the cached versions if they exist. cache_dir (`Union[str, os.PathLike]`, *optional*): Path to a directory where a downloaded pretrained model configuration is cached if the standard cache is not used. proxies (`Dict[str, str]`, *optional*): A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request. output_loading_info(`bool`, *optional*, defaults to `False`): Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages. local_files_only (`bool`, *optional*, defaults to `False`): Whether to only load local model weights and configuration files or not. If set to `True`, the model won't be downloaded from the Hub. token (`str` or *bool*, *optional*): The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from `diffusers-cli login` (stored in `~/.huggingface`) is used. revision (`str`, *optional*, defaults to `"main"`): The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier allowed by Git. custom_revision (`str`, *optional*, defaults to `"main"`): The specific model version to use. It can be a branch name, a tag name, or a commit id similar to `revision` when loading a custom pipeline from the Hub. It can be a 🤗 Diffusers version when loading a custom pipeline from GitHub, otherwise it defaults to `"main"` when loading from the Hub. mirror (`str`, *optional*): Mirror source to resolve accessibility issues if you’re downloading a model in China. We do not guarantee the timeliness or safety of the source, and you should refer to the mirror site for more information. device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*): A map that specifies where each submodule should go. It doesn’t need to be defined for each parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the same device. Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For more information about each option see [designing a device map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map). max_memory (`Dict`, *optional*): A dictionary device identifier for the maximum memory. Will default to the maximum memory available for each GPU and the available CPU RAM if unset. offload_folder (`str` or `os.PathLike`, *optional*): The path to offload weights if device_map contains the value `"disk"`. offload_state_dict (`bool`, *optional*): If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True` when there is some disk offload. low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`): Speed up model loading only loading the pretrained weights and not initializing the weights. This also tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model. Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this argument to `True` will raise an error. use_safetensors (`bool`, *optional*, defaults to `None`): If set to `None`, the safetensors weights are downloaded if they're available **and** if the safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors weights. If set to `False`, safetensors weights are not loaded. kwargs (remaining dictionary of keyword arguments, *optional*): Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline class). The overwritten components are passed directly to the pipelines `__init__` method. See example below for more information. variant (`str`, *optional*): Load weights from a specified variant filename such as `"fp16"` or `"ema"`. This is ignored when loading `from_flax`. To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with `hf auth login`. Examples: ```py >>> from diffusers import AutoPipelineForText2Image >>> pipeline = AutoPipelineForText2Image.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5") >>> image = pipeline(prompt).images[0] ``` cache_dirNforce_downloadFproxiestokenlocal_files_onlyrevisionrrrrrr _class_nameControlPipelinePipeline controlnetControlNetUnionPipelineControlNetPipeline enable_pag PAGPipeline)pop load_config isinstancerreplacerqrrfrom_pretrained)clspretrained_model_or_pathrrrrrrrload_config_kwargsconfigorig_class_name to_replacertext_2_image_clss ryrz)AutoPipelineForText2Image.from_pretrained*sddJJ{D1 $4e<**Y- 7D)!::&8%@::j$/#, 0  !!9P=OP /  /*J#J 6 !&.0DE"("7"?"? Le"f"("7"?"? L`"a 6 !L1J"1"9"9*m"T*+Lo^1&1&1///0HSFSSr{c Pt|j}|jj}t t |}d|vr|dPd|jvrdnd}t t |jj ddj |d|z}n*t t |jj dd}d|vrx|jd}|r;t t |jj ddj dd}n*t t |jj dd}|j|\}} |jd d} |D cic]} | |vs| |j| } } |jjD cic]\} } | |vr| | vr| |j| }} } | D cic]} | |vs| |j| }} |jD cic]\} } | | vr | |vr| || }} } |jD cgc]&} | jd r| d d| vr | d d|vr| d d(}} |D]} |jd | || <i| |||}|jD cic]&\} } | |vr| jd rdnd | || (}} } t|t|jz t|jz }t|d kDrOt!d |d|dtt#| jt#|jzd|di|}|j%| |j$di||Scc} wcc} } wcc} wcc} } wcc} wcc} } w)a> Instantiates a text-to-image Pytorch diffusion pipeline from another instantiated diffusion pipeline class. The from_pipe() method takes care of returning the correct pipeline class instance by finding the text-to-image pipeline linked to the pipeline class using pattern matching on pipeline class name. All the modules the pipeline contains will be used to initialize the new pipeline without reallocating additional memory. The pipeline is set in evaluation mode (`model.eval()`) by default. Parameters: pipeline (`DiffusionPipeline`): an instantiated `DiffusionPipeline` object ```py >>> from diffusers import AutoPipelineForText2Image, AutoPipelineForImage2Image >>> pipe_i2i = AutoPipelineForImage2Image.from_pretrained( ... "stable-diffusion-v1-5/stable-diffusion-v1-5", requires_safety_checker=False ... ) >>> pipe_t2i = AutoPipelineForText2Image.from_pipe(pipe_i2i) >>> image = pipe_t2i(prompt).images[0] ``` rNPAGrr ControlNetr _name_or_path_r r Pipeline expected , but only were passedr)dictrrrsrqrrrr_get_signature_keys componentsr~keys startswithset_optional_componentslenrlistregister_to_config)rrroriginal_configoriginal_cls_namerrrexpected_modulesoptional_kwargspretrained_model_name_or_pathkpassed_class_objvoriginal_class_objpassed_pipe_kwargsoriginal_pipe_kwargsadditional_pipe_kwargstext_2_image_kwargsunused_original_configmissing_modulesmodels ry from_pipez#AutoPipelineForText2Image.from_pipesg:x/$..77++LN_` 6 !l#/.37G7P7P.P]V` #25$--55lBGOOPZ\hku\uv$ $35$--55lBG$ 6 !L1J#25$--55eR@HHUbc$ $35$--55eR@$ -=,P,PQa,b)/(7(;(;OT(R%7GV!v+Avzz!},VV!++113 1$$2B)B x""1% %  9HW11PV;aA.WW(--/ 1O#1C(C q! !  %))+" ||C QqrUo%=!AB%OaBa abE" " ( CA&5&9&9AaS'&B  # Cw!1v5GvK]vauv (--/" 1++\\#&rC0 4oa6H H" "   !C(8(M(M$N NQTUhUmUmUoQp p   ! #,-Z8H7IUXY]^n^s^s^uYvy}QVVXzYZYVZU[[gh !7#67   /L M   :#9: _W X " " s0% N/N""N  NN>N++N+N"N rs __module__ __qualname____doc__ config_namer classmethodrrrrr{ryrrsJ %K TTTTlqqr{rcDeZdZdZdZdZeedZedZ y)AutoPipelineForImage2Imagea; [`AutoPipelineForImage2Image`] is a generic pipeline class that instantiates an image-to-image pipeline class. The specific underlying pipeline class is automatically selected from either the [`~AutoPipelineForImage2Image.from_pretrained`] or [`~AutoPipelineForImage2Image.from_pipe`] methods. This class cannot be instantiated using `__init__()` (throws an error). Class attributes: - **config_name** (`str`) -- The configuration filename that stores the class and module names of all the diffusion pipeline's components. rct|jjd|jjd|jjdrrrs ryrz#AutoPipelineForImage2Image.__init__Irr{c |jdd}|jdd}|jdd}|jdd}|jdd}|jdd}||||||d } |j|fi| } | d } d | vrd } n d | vrd } nd} d|vr>t|dtr| j | d| z} n| j | d| z} d|vr(|jd} | r| j | d| z} | d k(r| j | d} t t | }i| |}|j|fi|S)uK Instantiates a image-to-image Pytorch diffusion pipeline from pretrained pipeline weight. The from_pretrained() method takes care of returning the correct pipeline class instance by: 1. Detect the pipeline class of the pretrained_model_or_path based on the _class_name property of its config object 2. Find the image-to-image pipeline linked to the pipeline class using pattern matching on pipeline class name. If a `controlnet` argument is passed, it will instantiate a [`StableDiffusionControlNetImg2ImgPipeline`] object. The pipeline is set in evaluation mode (`model.eval()`) by default. If you get the error message below, you need to finetune the weights for your downstream task: ``` Some weights of UNet2DConditionModel were not initialized from the model checkpoint at stable-diffusion-v1-5/stable-diffusion-v1-5 and are newly initialized because the shapes did not match: - conv_in.weight: found shape torch.Size([320, 4, 3, 3]) in the checkpoint and torch.Size([320, 9, 3, 3]) in the model instantiated You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. ``` Parameters: pretrained_model_or_path (`str` or `os.PathLike`, *optional*): Can be either: - A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline hosted on the Hub. - A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights saved using [`~DiffusionPipeline.save_pretrained`]. torch_dtype (`str` or `torch.dtype`, *optional*): Override the default `torch.dtype` and load the model with another dtype. force_download (`bool`, *optional*, defaults to `False`): Whether or not to force the (re-)download of the model weights and configuration files, overriding the cached versions if they exist. cache_dir (`Union[str, os.PathLike]`, *optional*): Path to a directory where a downloaded pretrained model configuration is cached if the standard cache is not used. proxies (`Dict[str, str]`, *optional*): A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request. output_loading_info(`bool`, *optional*, defaults to `False`): Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages. local_files_only (`bool`, *optional*, defaults to `False`): Whether to only load local model weights and configuration files or not. If set to `True`, the model won't be downloaded from the Hub. token (`str` or *bool*, *optional*): The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from `diffusers-cli login` (stored in `~/.huggingface`) is used. revision (`str`, *optional*, defaults to `"main"`): The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier allowed by Git. custom_revision (`str`, *optional*, defaults to `"main"`): The specific model version to use. It can be a branch name, a tag name, or a commit id similar to `revision` when loading a custom pipeline from the Hub. It can be a 🤗 Diffusers version when loading a custom pipeline from GitHub, otherwise it defaults to `"main"` when loading from the Hub. mirror (`str`, *optional*): Mirror source to resolve accessibility issues if you’re downloading a model in China. We do not guarantee the timeliness or safety of the source, and you should refer to the mirror site for more information. device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*): A map that specifies where each submodule should go. It doesn’t need to be defined for each parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the same device. Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For more information about each option see [designing a device map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map). max_memory (`Dict`, *optional*): A dictionary device identifier for the maximum memory. Will default to the maximum memory available for each GPU and the available CPU RAM if unset. offload_folder (`str` or `os.PathLike`, *optional*): The path to offload weights if device_map contains the value `"disk"`. offload_state_dict (`bool`, *optional*): If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True` when there is some disk offload. low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`): Speed up model loading only loading the pretrained weights and not initializing the weights. This also tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model. Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this argument to `True` will raise an error. use_safetensors (`bool`, *optional*, defaults to `None`): If set to `None`, the safetensors weights are downloaded if they're available **and** if the safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors weights. If set to `False`, safetensors weights are not loaded. kwargs (remaining dictionary of keyword arguments, *optional*): Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline class). The overwritten components are passed directly to the pipelines `__init__` method. See example below for more information. variant (`str`, *optional*): Load weights from a specified variant filename such as `"fp16"` or `"ema"`. This is ignored when loading `from_flax`. To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with `hf auth login`. Examples: ```py >>> from diffusers import AutoPipelineForImage2Image >>> pipeline = AutoPipelineForImage2Image.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5") >>> image = pipeline(prompt, image).images[0] ``` rNrFrrrrrrImg2ImgImg2ImgPipelinerrrControlNetUnionrrrControlImg2ImgPipeline)rrrrrrqrur)rrrrrrrrrrrrrrimage_2_image_clss ryrz*AutoPipelineForImage2Image.from_pretrainedPsfJJ{D1 $4e<**Y- 7D)!::&8%@::j$/#, 0  !!9P=OP /  '*J / 1*J#J 6 !&.0DE"1"9"9*FWZdFd"e"1"9"9*lU_F_"` 6 !L1J"1"9"9*ejFX"Y * *-55jBZ[O+,NP_`1&1&10 001ITVTTr{c Vt|j}|jj}t t |}d|vr|dSd}d|jvrd|z}t t |jj ddj |d|z}n*t t |jj dd}d|vrx|jd}|r;t t |jj ddj dd}n*t t |jj dd}|j|\}} |jd d} |D cic]} | |vs| |j| } } |jjD cic]\} } | |vr| | vr| |j| }} } | D cic]} | |vs| |j| }} |jD cic]\} } | | vr | |vr| || }} } |jD cgc]&} | jd r| d d| vr | d d|vr| d d(}} |D]} |jd | || <i| |||}|jD cic]&\} } | |vr| jd rdnd | || (}} } t|t|jz t|jz }t|d kDrOt!d |d|dtt#| jt#|jzd|di|}|j%| |j$di||Scc} wcc} } wcc} wcc} } wcc} wcc} } w)aZ Instantiates a image-to-image Pytorch diffusion pipeline from another instantiated diffusion pipeline class. The from_pipe() method takes care of returning the correct pipeline class instance by finding the image-to-image pipeline linked to the pipeline class using pattern matching on pipeline class name. All the modules the pipeline contains will be used to initialize the new pipeline without reallocating additional memory. The pipeline is set in evaluation mode (`model.eval()`) by default. Parameters: pipeline (`DiffusionPipeline`): an instantiated `DiffusionPipeline` object Examples: ```py >>> from diffusers import AutoPipelineForText2Image, AutoPipelineForImage2Image >>> pipe_t2i = AutoPipelineForText2Image.from_pretrained( ... "stable-diffusion-v1-5/stable-diffusion-v1-5", requires_safety_checker=False ... ) >>> pipe_i2i = AutoPipelineForImage2Image.from_pipe(pipe_t2i) >>> image = pipe_i2i(prompt, image).images[0] ``` rNrrrrrPAGImg2ImgPipelinerrr rrrrrrr)rrrrsrqrurrrrr~rrrrrrrr)rrrrrrrrrrrrrrrrrrimage_2_image_kwargsrrrs ryrz$AutoPipelineForImage2Image.from_pipesp>x/$..77,,NPab 6 !l#/. -666!&!3J$36%..66|RHPP"L:$=%!%46%..66|RH%! 6 !L1J$36%..66ubAIIJ[]qr%! %46%..66ubA%! ->,Q,QRc,d)/(7(;(;OT(R%7GV!v+Avzz!},VV!++113 1$$2B)B x""1% %  9HW11PV;aA.WW(--/ 1O#1C(C q! !  %))+" ||C QqrUo%=!AB%OaBa abE" " ( CA&5&9&9AaS'&B  # C x"2w6HwL^wbvw (--/" 1,,\\#&rC0 4oa6H H" "   !C(9(N(N$O ORUVjVoVoVqRr r   ! #-.j9I8J+VYZ^_o_t_t_vZwz~@R@W@W@Y{Z[ZW[V\\hi "9$89   /L M   :#9: _W X " " s0( N 2N %"N NNN.+N +N%Nrrr{ryrr7sJ %K _U_UBwwr{rcDeZdZdZdZdZeedZedZ y)AutoPipelineForInpaintinga4 [`AutoPipelineForInpainting`] is a generic pipeline class that instantiates an inpainting pipeline class. The specific underlying pipeline class is automatically selected from either the [`~AutoPipelineForInpainting.from_pretrained`] or [`~AutoPipelineForInpainting.from_pipe`] methods. This class cannot be instantiated using `__init__()` (throws an error). Class attributes: - **config_name** (`str`) -- The configuration filename that stores the class and module names of all the diffusion pipeline's components. rct|jjd|jjd|jjdrrrs ryrz"AutoPipelineForInpainting.__init__rr{c |jdd}|jdd}|jdd}|jdd}|jdd}|jdd}||||||d } |j|fi| } | d } d | vrd } n d | vrd } nd} d|vr>t|dtr| j | d| z} n| j | d| z} d|vr(|jd} | r| j | d| z} | d k(r| j | d} t t | }i| |}|j|fi|S)uT Instantiates a inpainting Pytorch diffusion pipeline from pretrained pipeline weight. The from_pretrained() method takes care of returning the correct pipeline class instance by: 1. Detect the pipeline class of the pretrained_model_or_path based on the _class_name property of its config object 2. Find the inpainting pipeline linked to the pipeline class using pattern matching on pipeline class name. If a `controlnet` argument is passed, it will instantiate a [`StableDiffusionControlNetInpaintPipeline`] object. The pipeline is set in evaluation mode (`model.eval()`) by default. If you get the error message below, you need to finetune the weights for your downstream task: ``` Some weights of UNet2DConditionModel were not initialized from the model checkpoint at stable-diffusion-v1-5/stable-diffusion-v1-5 and are newly initialized because the shapes did not match: - conv_in.weight: found shape torch.Size([320, 4, 3, 3]) in the checkpoint and torch.Size([320, 9, 3, 3]) in the model instantiated You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. ``` Parameters: pretrained_model_or_path (`str` or `os.PathLike`, *optional*): Can be either: - A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline hosted on the Hub. - A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights saved using [`~DiffusionPipeline.save_pretrained`]. torch_dtype (`str` or `torch.dtype`, *optional*): Override the default `torch.dtype` and load the model with another dtype. force_download (`bool`, *optional*, defaults to `False`): Whether or not to force the (re-)download of the model weights and configuration files, overriding the cached versions if they exist. cache_dir (`Union[str, os.PathLike]`, *optional*): Path to a directory where a downloaded pretrained model configuration is cached if the standard cache is not used. proxies (`Dict[str, str]`, *optional*): A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request. output_loading_info(`bool`, *optional*, defaults to `False`): Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages. local_files_only (`bool`, *optional*, defaults to `False`): Whether to only load local model weights and configuration files or not. If set to `True`, the model won't be downloaded from the Hub. token (`str` or *bool*, *optional*): The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from `diffusers-cli login` (stored in `~/.huggingface`) is used. revision (`str`, *optional*, defaults to `"main"`): The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier allowed by Git. custom_revision (`str`, *optional*, defaults to `"main"`): The specific model version to use. It can be a branch name, a tag name, or a commit id similar to `revision` when loading a custom pipeline from the Hub. It can be a 🤗 Diffusers version when loading a custom pipeline from GitHub, otherwise it defaults to `"main"` when loading from the Hub. mirror (`str`, *optional*): Mirror source to resolve accessibility issues if you’re downloading a model in China. We do not guarantee the timeliness or safety of the source, and you should refer to the mirror site for more information. device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*): A map that specifies where each submodule should go. It doesn’t need to be defined for each parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the same device. Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For more information about each option see [designing a device map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map). max_memory (`Dict`, *optional*): A dictionary device identifier for the maximum memory. Will default to the maximum memory available for each GPU and the available CPU RAM if unset. offload_folder (`str` or `os.PathLike`, *optional*): The path to offload weights if device_map contains the value `"disk"`. offload_state_dict (`bool`, *optional*): If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True` when there is some disk offload. low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`): Speed up model loading only loading the pretrained weights and not initializing the weights. This also tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model. Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this argument to `True` will raise an error. use_safetensors (`bool`, *optional*, defaults to `None`): If set to `None`, the safetensors weights are downloaded if they're available **and** if the safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors weights. If set to `False`, safetensors weights are not loaded. kwargs (remaining dictionary of keyword arguments, *optional*): Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline class). The overwritten components are passed directly to the pipelines `__init__` method. See example below for more information. variant (`str`, *optional*): Load weights from a specified variant filename such as `"fp16"` or `"ema"`. This is ignored when loading `from_flax`. To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with `hf auth login`. Examples: ```py >>> from diffusers import AutoPipelineForInpainting >>> pipeline = AutoPipelineForInpainting.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5") >>> image = pipeline(prompt, image=init_image, mask_image=mask_image).images[0] ``` rNrFrrrrrrInpaintInpaintPipelinerrrrrrrControlInpaintPipeline)rrrrrrqrwr)rrrrrrrrrrrrrrinpainting_clss ryrz)AutoPipelineForInpainting.from_pretrainedsdJJ{D1 $4e<**Y- 7D)!::&8%@::j$/#, 0  !!9P=OP /  '*J / 1*J#J 6 !&.0DE"1"9"9*FWZdFd"e"1"9"9*lU_F_"` 6 !L1J"1"9"9*ejFX"Y * *-55jBZ[O()GY1&1&1-~--.FQ&QQr{c &t|j}|jj}t t |}d|vrj|d;t t |jj ddj dd}n*t t |jj dd}d|vrx|jd}|r;t t |jj ddj dd }n*t t |jj d d}|j|\}}|jd d} |D cic]} | |vs| |j| } } |jjD cic]\} } | |vr| | vr| |j| } } } |D cic]} | |vs| |j| }} |jD cic]\} } | |vr | |vr| || }} } |jD cgc]&} | jd r| d d|vr | d d|vr| d d(}} |D]} |jd | || <i| | ||}|jD cic]&\} } | |vr| jd rdnd | || (}} } t|t|jz t|jz }t|d kDrOt!d|d|dtt#| jt#| jzd|di|}|j%| |j$di||Scc} wcc} } wcc} wcc} } wcc} wcc} } w)aj Instantiates a inpainting Pytorch diffusion pipeline from another instantiated diffusion pipeline class. The from_pipe() method takes care of returning the correct pipeline class instance by finding the inpainting pipeline linked to the pipeline class using pattern matching on pipeline class name. All the modules the pipeline class contain will be used to initialize the new pipeline without reallocating additional memory. The pipeline is set in evaluation mode (`model.eval()`) by default. Parameters: pipeline (`DiffusionPipeline`): an instantiated `DiffusionPipeline` object Examples: ```py >>> from diffusers import AutoPipelineForText2Image, AutoPipelineForInpainting >>> pipe_t2i = AutoPipelineForText2Image.from_pretrained( ... "DeepFloyd/IF-I-XL-v1.0", requires_safety_checker=False ... ) >>> pipe_inpaint = AutoPipelineForInpainting.from_pipe(pipe_t2i) >>> image = pipe_inpaint(prompt, image=init_image, mask_image=mask_image).images[0] ``` rNrrrControlNetInpaintPipelinerrPAGInpaintPipelinerrr rrrrrrr)rrrrsrqrwrrrrr~rrrrrrrr)rrrrrrrrrrrrrrrrrinpainting_kwargsrrrs ryrz#AutoPipelineForInpainting.from_pipe'sI<x/$..77))GIZ[ 6 !l#/!02"++33L"EMM)+F""12"++334OQbc" 6 !L1J!02"++33E2>FFGXZno" "12"++334HJ[\" -;,N,N~,^)/(7(;(;OT(R%7GV!v+Avzz!},VV!++113 1$$2B)B x""1% %  9HW11PV;aA.WW(--/ 1O#1C(C q! !  %))+" ||C QqrUo%=!AB%OaBa abE" " ( CA&5&9&9AaS'&B  # Cu/t3EtI[t_st (--/" 1))\\#&rC0 4oa6H H" "   !C(K(K$L LsSdSiSiSkOl l   ! #N+:6F5G{SVW[\l\q\q\sWtw{}O}T}T}VxWXWTXSYYef 3!23   /L M   :#9: _W X " " s0 M2M2 "M76 M=M=)N+N>+N Nrrr{ryrrnsJ %K \R\R|ssr{rN)T) collectionsrhuggingface_hub.utilsrconfiguration_utilsrmodels.controlnetsrutilsr aura_flowr rfr rgr rhr rrrrrrrrrrrcontrolnet_sd3rr deepfloyd_ifrrrrcrrrr r!r"r#r$r%r& hunyuanditr'r[r(r)r*r+r,r- kandinsky2_2r.r/r0r1r2r3r]r4r5latent_consistency_modelsr6r7rdr8rer9pagr:r;r<r=r>r?r@rArBrCrDrErFrGrH pixart_alpharIrJrarKstable_cascaderLrMstable_diffusionrNrOrPstable_diffusion_3rQrRrSstable_diffusion_xlrTrUrVr^rWrXrrrurwrortrvrlrirjrkr}rzrboolrqrrrrr{ryr s $6-5.'"*?   NM   +Fl"$"CW   N%0& 45&  9:& 78& "#>? & z & &' & -.& /0& 45& )*& '(IJ& *+NO& 01YZ& )*LM& 12& 12!&" ./#&$ ,-%&& ,-'&( )&* _%+&, !;<-&. +,PQ/&0 #$@A1&2 ./UV3&4 345&6 %&7&8 9&: ,-;&< 23=&> ,-?&@ >"A&B O$C&D >"E&F )*G&H %&I&J 45K&(%!T&1 ;<  @A >? !#EF  ! 67 ;< 01 &(PQ !BC )+UV /1`a "$GH -/\] 56 $% 9: 34 ,-'&"0"- ;<  @A >? #$ 67 ;< &(PQ *,WX )+UV /1`a (*VW "$GH $% 9: 34 !BC!"*.9 '( ,- 01 01 .*/: ./ 34/+ +6 ./ 34+'=&2@%h/6G%l33H&x0&&"./+  v" qD qc cL ttn m mr{