pyiu%SrSSKrSSKrSSKJr SSKJr SSKJr SSK J r J r SSK J r SS KJrJrJrJrJrJrJrJrJr SS KJr S S KJr S S KJr S SKJrJ r J!r!J"r" \RF"\$5r%/SQr&\(a\"5r\\'\(\'\'S-44\)S'GOR\"/SSSS.4PSSSS.4PSSSS.4PSSSS.4PSSSS.4PSSSS.4PS S!S"S.4PS#S$S%S.4PS&S'S(S.4PS)S*S+S.4PS,S*S+S.4PS-S.S/S.4PS0S1S2S.4PS3S4S5S.4PS6S!S"S.4PS7S8S9S.4PS:S;SS?04PS@SASB04PSCSDSES.4PSFS'S(S.4PSGSSS.4PSHS8S9S.4PSIS$S%S.4PSJSSS.4PSKSLSMS.4PSNSSS.4PSOS8S9S.4PSPS!S"S.4PSQSSS.4PSRSSS.4PSSSSS.4PSTSUSVS.4PSWSXSYS.4PSZS[S\S.4PS]SLSMS.4PS^S_S`S.4PSaSSS.4PSbS!S"S.4PScS[S\S.4PSdSSS.4PSeSfSgS.4PShSiSjS.4PSkSlSmS.4PSnS$S%S.4PSoS8S9S.4PSpS>Sq04PSrS>Ss04PStSuSvS.4PSwSSS.4PSxS'S(S.4PSyS'S(S.4PSzS'S(S.4PS{S'S(S.4PS|S'S(S.4PS}S'S(S.4PS~S*S+S.4PSS*S+S.4PSS>S?04PSS>S04PSS>S04PSS>S04PSS>S04PSSSS.4PSSSS.4PSS>S04PSS!S"S.4PSS!S"S.4PSS!S"S.4PSSSS.4PSSSS.4PSSlSmS.4PSSSS.4PSSAS04PSS!S"S.4PSSSS.4PSSSS.4PSSAS04PSSSS.4PSS!S"S.4PSS!S"S.4PSS>S04PSSSS.4P5r*\RV"\*5 \"\\5r,S\'4Sjr-SS\'\R\-S\'\R\-S-S\/S\(\'\'4S-S\/\'-S-S\'S-S\/4Sjjr0S\'S-S\/S-S\'S-S\'4Sjr1Sr2S\'S\(S-4Sjr3SSjr4Sr5\"SS9"SS55r6SS/r7g)zAutoImageProcessor class.N) OrderedDict) TYPE_CHECKING)PreTrainedConfig)get_class_from_dynamic_moduleresolve_trust_remote_code)ImageProcessingMixin) CONFIG_NAMEIMAGE_PROCESSOR_NAMEPROCESSOR_NAME cached_fileis_timm_config_dictis_timm_local_checkpointis_torchvision_availableloggingsafe_load_json_file)requires)_LazyAutoMapping)IMAGE_PROCESSOR_MAPPING_NAMES)CONFIG_MAPPING_NAMES AutoConfigmodel_type_to_module_name!replace_list_option_in_docstrings)ChameleonImageProcessorFlavaImageProcessorIdefics3ImageProcessorSmolVLMImageProcessorraimv2CLIPImageProcessorCLIPImageProcessorPil torchvisionpilaimv2_vision_modelalignEfficientNetImageProcessorEfficientNetImageProcessorPilaltclip aya_visionGotOcr2ImageProcessorGotOcr2ImageProcessorPilzblip-2BlipImageProcessorBlipImageProcessorPilclipsegViTImageProcessorViTImageProcessorPilcolpaliSiglipImageProcessorSiglipImageProcessorPilcolqwen2Qwen2VLImageProcessorQwen2VLImageProcessorPil convnextv2ConvNextImageProcessorConvNextImageProcessorPilcvtzdata2vec-visionBeitImageProcessorBeitImageProcessorPildeimv2RTDetrImageProcessorRTDetrImageProcessorPildepth_anythingDPTImageProcessorDPTImageProcessorPildinatdinov2BitImageProcessorBitImageProcessorPilz donut-swinDonutImageProcessorDonutImageProcessorPiledgetamr#Sam2ImageProcessoremu3r$Emu3ImageProcessor eomt_dinov3EomtImageProcessorEomtImageProcessorPil exaone4_5 florence2focalnetgemma3ngitgranite4_visionLlavaNextImageProcessorLlavaNextImageProcessorPilgroupvithieraijepa instructblipinternvlzkosmos-2z kosmos-2.5Kosmos2_5ImageProcessorKosmos2_5ImageProcessorPil layoutxlmLayoutLMv2ImageProcessorLayoutLMv2ImageProcessorPil lighton_ocrPixtralImageProcessorPixtralImageProcessorPilllava_next_videolw_detrDeformableDetrImageProcessorDeformableDetrImageProcessorPil metaclip_2zmgp-strmistral3mlcdzmm-grounding-dinoGroundingDinoImageProcessorGroundingDinoImageProcessorPil mobilevitv2MobileViTImageProcessorMobileViTImageProcessorPilz omdet-turboDetrImageProcessorDetrImageProcessorPil paligemmapixiopp_ocrv5_mobile_detPPOCRV5ServerDetImageProcessorpp_ocrv5_mobile_recPPOCRV5ServerRecImageProcessorpvt_v2PvtImageProcessorPvtImageProcessorPil qianfan_ocr qwen2_5_omni qwen2_5_vlqwen3_5 qwen3_5_moeqwen3_omni_moeqwen3_vlregnetresnet sam2_videosam3_lite_textSam3ImageProcessor sam3_trackersam3_tracker_video sam3_videosam_hqSamImageProcessorSamImageProcessorPil shieldgemma2Gemma3ImageProcessorGemma3ImageProcessorPilslanetSLANeXtImageProcessor swiftformerswinswinv2t5gemma2t5gemma2_encoderztable-transformer timesformerVideoMAEImageProcessorPilVideoMAEImageProcessor)r$r# timm_wrapperTimmWrapperImageProcessortrocrudopLayoutLMv3ImageProcessorLayoutLMv3ImageProcessorPilupernetSegformerImageProcessorSegformerImageProcessorPil video_llavaVideoLlavaImageProcessorvipllavavit_maevit_msnvivitVivitImageProcessorxclip class_namec@US:XaSSKJn U$[RR 5HHnUR 5H1n[ U[ 5(dM[USS5U:XdM-Us s $ MJ [R"5HKupEXR 5;dM[U5n[R"SU3S5n[Xp5s $ [R"S5n[X5(a [X5$g![a Mf=f) zResolve an image processor class name to its class. Handles both base names (e.g. CLIPImageProcessor) and PIL backend names (e.g. CLIPImageProcessorPil). No recursion needed since names are direct.BaseImageProcessorFastr)TorchvisionBackend__name__N.ztransformers.models transformers)image_processing_backendsrIMAGE_PROCESSOR_MAPPING_extra_contentvalues isinstancetypegetattrritemsr importlib import_moduleAttributeErrorhasattr) rrmappingextractor_class model_typeextractors_dict module_namemodule main_modules k/home/mande/repo/quber/.venv/lib/python3.13/site-packages/transformers/models/auto/image_processing_auto.py#get_image_processor_class_from_namers--C!!+99@@B&~~/O/400W_jZ^5_cm5m&& 0C (E'J'J'L# //1 13J?K,,q ->@UVF v22 (M)).9K{''{// "  s D DDpretrained_model_name_or_path cache_dirforce_downloadproxiestokenrevisionlocal_files_onlyc [U[UUUUUUSSS9 n[U[UUUUUUSSS9 n U (dU(d[R S5 0$0n Ub[ U5n SU ;aU Sn U bU c [ U 5n U $)aA Loads the image processor configuration from a pretrained model image processor configuration. Args: pretrained_model_name_or_path (`str` or `os.PathLike`): This can be either: - a string, the *model id* of a pretrained model configuration hosted inside a model repo on huggingface.co. - a path to a *directory* containing a configuration file saved using the [`~ProcessorMixin.save_pretrained`] method, e.g., `./my_model_directory/`. cache_dir (`str` or `os.PathLike`, *optional*): Path to a directory in which a downloaded pretrained model configuration should be cached if the standard cache should not be used. force_download (`bool`, *optional*, defaults to `False`): Whether or not to force to (re-)download the configuration files and override the cached versions if they exist. proxies (`dict[str, str]`, *optional*): A dictionary of proxy servers to use by protocol or endpoint, e.g., `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.` The proxies are used on each request. token (`str` or *bool*, *optional*): The token to use as HTTP bearer authorization for remote files. If `True`, will use the token generated when running `hf auth login` (stored in `~/.huggingface`). 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, since we use a git-based system for storing models and other artifacts on huggingface.co, so `revision` can be any identifier allowed by git. local_files_only (`bool`, *optional*, defaults to `False`): If `True`, will only try to load the image processor configuration from local files. Passing `token=True` is required when you want to use a private model. Returns: `Dict`: The configuration of the image processor. Examples: ```python # Download configuration from huggingface.co and cache. image_processor_config = get_image_processor_config("google-bert/bert-base-uncased") # This model does not have a image processor config so the result will be an empty dict. image_processor_config = get_image_processor_config("FacebookAI/xlm-roberta-base") # Save a pretrained image processor locally and you can reload its config from transformers import AutoImageProcessor image_processor = AutoImageProcessor.from_pretrained("google/vit-base-patch16-224-in21k") image_processor.save_pretrained("image-processor-test") image_processor_config = get_image_processor_config("image-processor-test") ```F) filenamerrrrrr _raise_exceptions_for_gated_repo%_raise_exceptions_for_missing_entriesz8Could not locate the image processor configuration file.image_processor)r r r loggerinfor) rrrrrrrkwargsresolved_processor_fileresolved_image_processor_fileimage_processor_dictprocessor_dicts rget_image_processor_configrsD*%%))..3 %0%%%))..3 %! )1H NO *,-DE  .#12C#D $05I5Q23PQ backenduse_fastbase_class_namereturncUb#[RS5 Uc U(aSOSnUcU[;ag[5(aS$S$U$)aResolve raw backend inputs to a concrete backend name ('torchvision' or 'pil'). Handles, in order: - Deprecated ``use_fast`` flag: warns and converts to an explicit backend string when no explicit backend is given. - Explicit backend string: returned as-is. - None resolution: forces 'pil' for processors in DEFAULT_TO_PIL_BACKEND_IMAGE_PROCESSORS (Lanczos interpolation, unsupported by torchvision); otherwise picks 'torchvision' when available, falling back to 'pil'. zThe `use_fast` parameter is deprecated and will be removed in a future version. Use `backend="torchvision"` instead of `use_fast=True`, or `backend="pil"` instead of `use_fast=False`.r#r$)r warning_once'DEFAULT_TO_PIL_BACKEND_IMAGE_PROCESSORSr)rrrs r_resolve_backendr4sW v  ?'/mUG E E 8 : :}EE NrcBU/UVs/sH o"U:wdM UPM sn-nUHvnURU5nUcM[U[5(aUnO [U5nUb[ USS5(aMSXA:wa[ R SUSUS35 Us $ gs snf)aY Load an image processor class from a backend-to-class mapping, with fallback. Tries the requested backend first, then the opposite standard backend, then any remaining backends. Works with both string class names and resolved class objects. Unavailable backends are detected via DummyObject: classes whose required libraries are missing are represented as DummyObject subclasses (is_dummy=True). When the torchvision backend is missing but a PIL variant exists, _LazyModule transparently returns the PIL class with its own warning, so _load_class_with_fallback naturally receives a usable class without extra gating. Args: mapping: dict mapping backend names (str) to class names (str) or class objects (type). backend: the preferred backend name (e.g. "torchvision", "pil"). Returns: The loaded class, or None if no class could be loaded. Nis_dummyFz Requested z+ backend is not available. Falling back to z backend.)getrrrrrr)rrkbackends_to_trybvalueprocessor_classs r_load_class_with_fallbackrOs&ig"Fgg1g"FFO  A =  eT " "#OA%HO  "goz5&Q&Q  <   *WI5`ab`ccl m n!$ )#Gs BBc`^^S[S[4Sjm[R"5H0n[ UU4SjUR555(dM.Us $ [ R R5H0n[ UU4SjUR555(dM.Us $ g)z Find the backend->class mapping that contains base_class_name in its values. Returns the mapping dict (including any custom registered backends) or None. namercUcg[U[5(aX:H$[U[5(a[USS5U:H$g)NFr)rstrrr)valrs r_value_matches9_find_mapping_for_image_processor.._value_matchessC ; c3  ;  c4 3 D1T9 9rc36># UHnT"UT5v M g7fN.0vrrs r 4_find_mapping_for_image_processor..sQ;Pa~a11;Pc36># UHnT"UT5v M g7frrrs rrrsL;Ka~a11;KrN)rboolrranyrr)r mapping_dictcontentrs` @r!_find_mapping_for_image_processorrys #$6<<> Q<;N;N;PQ Q Q ?+99@@B L7>>;KL L LNC rcx[U5nUcUUS-S.n[X15nUcU(a[US-5nU$)a  Load image processor class for a given backend. Uses the mapping from IMAGE_PROCESSOR_MAPPING when base_class_name is found in its values (so config overrides and custom backends are respected). Falls back to base+Pil convention for remote code / unknown processors. Pilr"Fast)rrr)rris_legacy_fastrrs r_load_backend_classr sQ0@G*"U* 0AO>=oPV>VW rc$[U[5(a:URU5=(d" [[ UR 555$[U[ [45(a%US:Xa[U5S:a USbUS$US$U$)abExtract the class reference string from an auto_map entry based on backend preference. Returns: A string that may be: - A simple class name (e.g. `"MyImageProcessor"`) - A Hub reference in the form `upstream_repo--path/to/file.py::ClassName`, where the part before `--` is the upstream repo ID (used for trust_remote_code resolution). r#rr) rdictrnextiterrlisttuplelen)auto_maprs r_resolve_auto_map_class_refrs|(D!!||G$ET(//2C-D(EE(T5M** m #H (9hqk>UA; { Or)vision)backendsc \rSrSrSrSr\\"\5S55r \ S S\ S-S\ S-S\ \ \ 4S-S \4S jj5rS rg) AutoImageProcessoria This is a generic image processor class that will be instantiated as one of the image processor classes of the library when created with the [`AutoImageProcessor.from_pretrained`] class method. This class cannot be instantiated directly using `__init__()` (throws an error). c[S5e)NzAutoImageProcessor is designed to be instantiated using the `AutoImageProcessor.from_pretrained(pretrained_model_name_or_path)` method.)OSError)selfs r__init__AutoImageProcessor.__init__s d  rcnURSS5nURSS5nURSS5nURSS5nSUS'SU;aURS5nO[U5(a[nO[n[R "U4SU0UD6upU RS S5n Sn S U RS 05;aU S S n U cZU cWU RS S5nUbURS S5n SU RS 05;aU S SnURS S5n U csU bUSLak[U[5(d[R"U4SU0UD6n[US S5n [US 5(aS UR ;aUR S n SnSnU bU R#S5nU(aU SSOU n[%XeU5nSnUb ['UUU5nU SLnUSL=(d [)U5[*;nU=(aC U=(d [-[*[)U5U5R.R1S5(+nU(a6[3U U5nSU;aUR5S5SOSn[7XqUUU5nU(aSU(aLU(dE[9WU40UD6nURSS5n UR;5 UR"U/UQ70UD6$UbUR"U/UQ70UD6$[)U5[*;a~[*[)U5n[-UU5nUbUR"U/UQ70UD6$UR=5VVs/sH unnUcM UPM nnn[?SSRAU535e[?SUS[S[S[SSRAS[B553 5e![ aNn [R "U4S[0UD6upO![ a U ef=f[U 5(dU eSn A GNSn A ff=fs snnf)a Instantiate one of the image processor classes of the library from a pretrained model vocabulary. The image processor class to instantiate is selected based on the `model_type` property of the config object (either passed as an argument or loaded from `pretrained_model_name_or_path` if possible), or when it's missing, by falling back to using pattern matching on `pretrained_model_name_or_path`: List options Params: pretrained_model_name_or_path (`str` or `os.PathLike`): This can be either: - a string, the *model id* of a pretrained image_processor hosted inside a model repo on huggingface.co. - a path to a *directory* containing a image processor file saved using the [`~image_processing_utils.ImageProcessingMixin.save_pretrained`] method, e.g., `./my_model_directory/`. - a path to a saved image processor JSON *file*, e.g., `./my_model_directory/preprocessor_config.json`. cache_dir (`str` or `os.PathLike`, *optional*): Path to a directory in which a downloaded pretrained model image processor should be cached if the standard cache should not be used. force_download (`bool`, *optional*, defaults to `False`): Whether or not to force to (re-)download the image processor files and override the cached versions if they exist. proxies (`dict[str, str]`, *optional*): A dictionary of proxy servers to use by protocol or endpoint, e.g., `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.` The proxies are used on each request. token (`str` or *bool*, *optional*): The token to use as HTTP bearer authorization for remote files. If `True`, will use the token generated when running `hf auth login` (stored in `~/.huggingface`). 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, since we use a git-based system for storing models and other artifacts on huggingface.co, so `revision` can be any identifier allowed by git. use_fast (`bool`, *optional*, defaults to `False`): **Deprecated**: Use `backend="torchvision"` instead. This parameter is kept for backward compatibility. Use a fast torchvision-based image processor if it is supported for a given model. If a fast image processor is not available for a given model, a normal numpy-based image processor is returned instead. backend (`str`, *optional*, defaults to `None`): The backend to use for image processing. Can be: - `None`: Automatically select the best available backend (torchvision if available, otherwise pil) - `"torchvision"`: Use Torchvision backend (GPU-accelerated, faster) - `"pil"`: Use PIL backend (portable, CPU-only) - Any custom backend name registered via `register()` method return_unused_kwargs (`bool`, *optional*, defaults to `False`): If `False`, then this function returns just the final image processor object. If `True`, then this functions returns a `Tuple(image_processor, unused_kwargs)` where *unused_kwargs* is a dictionary consisting of the key/value pairs whose keys are not image processor attributes: i.e., the part of `kwargs` which has not been used to update `image_processor` and is otherwise ignored. trust_remote_code (`bool`, *optional*, defaults to `False`): Whether or not to allow for custom models defined on the Hub in their own modeling files. This option should only be set to `True` for repositories you trust and in which you have read the code, as it will execute code present on the Hub on your local machine. image_processor_filename (`str`, *optional*, defaults to `"config.json"`): The name of the file in the model directory to use for the image processor config. kwargs (`dict[str, Any]`, *optional*): The values in kwargs of any keys which are image processor attributes will be used to override the loaded values. Behavior concerning key/value pairs whose keys are *not* image processor attributes is controlled by the `return_unused_kwargs` keyword parameter. Passing `token=True` is required when you want to use a private model. Examples: ```python >>> from transformers import AutoImageProcessor >>> # Download image processor from huggingface.co and cache. >>> image_processor = AutoImageProcessor.from_pretrained("google/vit-base-patch16-224-in21k") >>> # If image processor files are in a directory (e.g. image processor was saved using *save_pretrained('./test/saved_model/')*) >>> # image_processor = AutoImageProcessor.from_pretrained("./test/saved_model/") ```configNrrtrust_remote_codeT _from_autoimage_processor_filenameimage_processor_typerrfeature_extractor_typeFeatureExtractorImageProcessorAutoFeatureExtractorFrz transformers.z--r code_revisionz:Could not find image processor class. Available backends: z, z Unrecognized image processor in z2. Should have a `image_processor_type` key in its z of z3, or one of the following `model_type` keys in its z: c3$# UHov M g7frr)rcs rr5AutoImageProcessor.from_pretrained..s@jLiqLis)"poprr r r get_image_processor_dict Exceptionrrreplacerrrfrom_pretrainedrrrendswithrr rrr __module__ startswithrsplitrrregister_for_auto_classr ValueErrorjoinr)clsrinputsrrr backend_kwargrr  config_dict_initial_exceptionr!image_processor_auto_mapfeature_extractor_classfeature_extractor_auto_maprrrimage_processor_classhas_remote_codehas_local_codeexplicit_local_code class_ref upstream_repoimage_processor_mappingrr availables rr/"AutoImageProcessor.from_pretrainedsfHd+::j$/ 9d3 "JJ':DA#| & /'-zz2L'M $ %&C D D'2 $'; $ (1JJ-H`djNK +/EtL#' ;??:r#B B'2:'>?S'T $ ',D,L&1oo6NPT&U #&2'>'F'FGY[k'l$%R)HH-8-DE[-\*+E+M+MN`br+s( '-E-MQbfkQkf&677#331&7 $+63I4#P vz**/Cv/V+1??;O+P(  +1::6BN;I23B7OcO"=OL $  &$7R`$a !3$>.d:ed6lNe>e,1 ! n%>?VW[\bWc?dfm%n *ZZ021 34LgVI8< 8IIOOD1!4tM 9!.Racp!  09L$A)Mj$unt$u ! ?D1A ! 9 9 ;(889VjY_jcij j " .(889VjY_jcij j &\4 4&=d6l&K #$=>UW^$_ !$0,<<=Zn]cngmnn'>'D'D'FX'Ftq!!'FIXYZ^ZcZcdmZnYopq q./L.MN11E0Fd;-X((3}Btyy@jLi@j7j6k m  c ( (!5!N!N1"LW"[a" Q ('' ('{33''4 (^Ys6O" P10P1 P.!PP) PP))P.Nslow_image_processor_classfast_image_processor_classimage_processor_classesexist_okc`Uc0nUbXS'UbX#S'U(d [S5eU[R;a[UnURU5 UnSSKJn UR 5H+upxUcM [X5(aM[SUSU35e [RXUS 9 g) av Register a new image processor for this class. Args: config_class ([`PreTrainedConfig`]): The configuration corresponding to the model to register. slow_image_processor_class (`type`, *optional*): The PIL backend image processor class (deprecated, use `image_processor_classes={"pil": ...}`). fast_image_processor_class (`type`, *optional*): The Torchvision backend image processor class (deprecated, use `image_processor_classes={"torchvision": ...}`). image_processor_classes (`dict[str, type]`, *optional*): Dictionary mapping backend names to image processor classes. Allows registering custom backends. Example: `{"pil": MyPilProcessor, "torchvision": MyTorchvisionProcessor, "custom": MyCustomProcessor}` exist_ok (`bool`, *optional*, defaults to `False`): If `True`, allow overwriting existing registrations. Nr$r#zYou need to specify at least one image processor class. Use `image_processor_classes={'backend_name': ProcessorClass}` or the deprecated `slow_image_processor_class`/`fast_image_processor_class` parameters.r)BaseImageProcessorz#Image processor class for backend 'z/' must inherit from `BaseImageProcessor`. Got: )rL) r5rrupdateimage_processing_utilsrNr issubclassregister) config_classrIrJrKrLexisting_mappingrN backend_keyrs rrRAutoImageProcessor.registers2 # *&( #)51K.)59S 6&X  2AA A6|D   # #$; <&6 # A,C,I,I,K (K*:o3b3b 9+G+,.-L ((Ya(brr)NNNF)rr1 __qualname____firstlineno____doc__r classmethodrrr/ staticmethodrr rrrR__static_attributes__rrrrrs &'DEz Fz x3726:> 5c$(4K5c%)4K5c"&c4i4!7 5c  5c5crrr)NFNNNF)F)8rYros collectionsrtypingrconfiguration_utilsrdynamic_module_utilsrrrPr utilsr r r r rrrrrutils.import_utilsr auto_factoryr auto_mappingsrconfiguration_autorrrr get_loggerrrrrr __annotations__%MISSING_IMAGE_PROCESSOR_MAPPING_NAMESrOrrPathLikerrrrrr rr__all__rrrrls   # 4\:   +*8   H %+'NY]!;sDcDj4I/I#JZ-8Z &:CZ[ \Z !3GPg#h iZ &BKjk lZ (G`a bZ )@Icd eZ +CLgh iZ $%Z &e12 3'Z (,@I`a b)Z **AJde f+Z ,*>G^_ `-Z .)G`a b1Z 2$8AXY Z3Z 40IRn o p5Z 6)=F]^ _7Z 8&9BXY Z9Z :&9BXY Z;Z <-AJab c=Z >)@Icd e?Z @)=F]^ _AZ B+DMij kCZ D*DMjk lEZ F,CLfg hGZ H 1JSo!p qIZ J(FOpq rKZ L+?H_` aMZ N(;DZ[ \OZ P)@Icd eQZ R%9BYZ [SZ V$#@; UZ b,ENjk lcZ d,@I`a beZ f*@Ibc dgZ h&9BXY ZiZ j#]4T$U VkZ l#]4T$U VmZ n':CYZ [oZ p,CLfg hqZ r-DMgh isZ t+BKef guZ v(?Hbc dwZ x,CLfg hyZ z/FOij k{Z |)@Icd e}Z ~'?Hcd eZ @'?Hcd eAZ BM+?@ ACZ D /CD EEZ Fm-AB CGZ H"M3G#H IIZ JM+?@ AKZ L':CYZ [MZ N-CLef gOZ P '>? @QZ R,?H^_ `SZ T%8AWX YUZ V':CYZ [WZ X)?Hab cYZ Z 1GPi!j k[Z \!2FOf"g h]Z ^$?Phi j_Z `e%@A BaZ b&9BXY ZcZ d%?Hef geZ f(AJfg hgZ hU$>? @iZ j)=F]^ _kZ l(;DZ[ \mZ n(;DZ[ \oZ p}&;< =qZ r&:CZ[ \sZ \-)|"(()NO*+?A^_CF+/ %)#"l #&#4l R[[ 4'l l #s(^d " l #:  l Dj l l ^cDjD4KRUX\R\ad6'Tstd{4,& ;BcBc BcJ %&: ;r