+h8SSKrSSKrSSKrSSKrSSKrSSKJrJ r SSK J r SSK J r Jr SSKJrJr \"\5r\R(R+S5SLr\R(R+S5SLr\R(R+S 5SLr\(a SSKrSS KJr O\"S 5e\(aSS KJr SS KJ r J!r! O\"S5e\(a SSK"J#r# SSK$J%r% O\"S5eSSjr&SSjr'Sr(SSjr)SSjr*Sr+g)N)ImageImageOps)InterpolationMode) normalizeresize) get_logger load_image insightfaceconsisid_eva_clipfacexlib) FaceAnalysiszPinsightface is not available. Please install it using 'pip install insightface'.)create_model_and_transforms)OPENAI_DATASET_MEANOPENAI_DATASET_STDz\consisid_eva_clip is not available. Please install it using 'pip install consisid_eva_clip'.)init_parsing_model)FaceRestoreHelperzJfacexlib is not available. Please install it using 'pip install facexlib'.cURSSup#[X#5U::aU$U[X#5- n[X$-5n[X4-5n[R"XU4[R S9nU$)at Resize the input image to a specified long edge while maintaining aspect ratio. Args: image (numpy.ndarray): Input image (H x W x C or H x W). resize_long_edge (int): The target size for the long edge of the image. Default is 768. Returns: numpy.ndarray: Resized image with the long edge matching `resize_long_edge`, while maintaining the aspect ratio. N) interpolation)shapemaxintcv2rINTER_LANCZOS4)imageresize_long_edgehwks e/home/james-whalen/.local/lib/python3.13/site-packages/diffusers/pipelines/consisid/consisid_utils.pyresize_numpy_image_longr"'sj ;;r?DA 1y$$ 3q9$A AE A AE A JJu!fC4F4F GE LcSn[U[5(aUVs/sH oC"XAU5PM sn$U"XU5$s snf)a/Numpy array to tensor. Args: imgs (list[ndarray] | ndarray): Input images. bgr2rgb (bool): Whether to change bgr to rgb. float32 (bool): Whether to change to float32. Returns: list[tensor] | tensor: Tensor images. If returned results only have one element, just return tensor. cBURSS:XaMU(aFURS:XaURS5n[R"U[R 5n[ R"URSSS55nU(aUR5nU$)Nrrfloat64float32r) rdtypeastypercvtColor COLOR_BGR2RGBtorch from_numpy transposefloat)imgbgr2rgbr's r! _totensorimg2tensor.._totensorKst 99Q<1 yyI%jj+,,sC$5$56Cs}}Q156 ))+C r#) isinstancelist)imgsr2r'r3r1s r! img2tensorr8>sE$<@ADS #0DAA TG ,,Bs;cSUSS2SS24-SUSS2SS24--SUSS2SS24--nURSSSS5nU$) a Converts an RGB image to grayscale by applying the standard luminosity formula. Args: img (torch.Tensor): The input image tensor with shape (batch_size, channels, height, width). The image is expected to be in RGB format (3 channels). Returns: torch.Tensor: The grayscale image tensor with shape (batch_size, 3, height, width). The grayscale values are replicated across all three channels. gA`"?Nrr(gbX9?rgv/?r)repeat)r1xs r!to_grayr<Zs_ AqsF ec!QqS&k11EC1Q3K4GGA Aq!A Hr#c J^UR5 [R"U[R5n UR U 5n [ U 5S:a[ U SS9Sn U Sn U SnOSn SnURU 5 URSS 9 UcURSnUR5 [ UR5S:Xa [S 5eURSnU c&[RS 5 URU5n [ R""U 5R%Xg5n U R&S :XaU R)S5n U (a[+USS 9R)S5S- nUR%U5nUR-[/U/SQ/SQ55SmTR1S SS9m/SQn[3U4SjU55R55n[ R6"U5n[ R8"UU[;U55n[ R8"UUU5nOV[R"U [R5n[+USS 9R)S5S- nUR%U5nU=nn[=UUR>[@RB5n[/UX45nU"UR%U5SSSS9unn[ RD"USS S5n[ RF"UU5n[ RH"U U/SS9nUUUU4$)a Process face embeddings from an image, extracting relevant features such as face embeddings, landmarks, and parsed face features using a series of face detection and alignment tools. Args: face_helper_1: Face helper object (first helper) for alignment and landmark detection. clip_vision_model: Pre-trained CLIP vision model used for feature extraction. face_helper_2: Face helper object (second helper) for embedding extraction. eva_transform_mean: Mean values for image normalization before passing to EVA model. eva_transform_std: Standard deviation values for image normalization before passing to EVA model. app: Application instance used for face detection. device: Device (CPU or GPU) where the computations will be performed. weight_dtype: Data type of the weights for precision (e.g., `torch.float32`). image: Input image in RGB format with pixel values in the range [0, 255]. original_id_image: (Optional) Original image for feature extraction if `is_align_face` is False. is_align_face: Boolean flag indicating whether face alignment should be performed. Returns: Tuple: - id_cond: Concatenated tensor of Ante face embedding and CLIP vision embedding - id_vit_hidden: Hidden state of the CLIP vision model, a list of tensors. - return_face_features_image_2: Processed face features image after normalization and parsing. - face_kps: Keypoints of the face detected in the image. rcHUSSUSS- USSUSS- -$)Nbboxrrrr()r;s r!)process_face_embeddings..s=QvYq\AfIaL5PUVW]U^_`UadefldmnodpUp4qr#)key embeddingkpsNT)only_center_facezfacexlib align face failzMFailed to detect face using insightface. Extracting embedding with align facer()r2go@)g ףp= ?gv/?gCl?)gZd;O?gy&1?g?)dimkeepdim)r c3.># UH nTU:Hv M g7f)Nr@).0i parsing_outs r! *process_face_embeddings..s48a!8sF)return_all_features return_hiddenshuffler)rH)% clean_allrr+ COLOR_RGB2BGRgetlensorted read_imageget_face_landmarks_5all_landmarks_5align_warp_face cropped_faces RuntimeErrorloggerwarningget_featr-r.tondim unsqueezer8 face_parserargmaxsumbool ones_likewherer<r image_sizerBICUBICnormdivcat) face_helper_1clip_vision_model face_helper_2eva_transform_meaneva_transform_stdappdevice weight_dtyperoriginal_id_image is_align_face image_bgr face_infoid_ante_embeddingface_kps align_faceinputbg_labelbg white_imagereturn_face_features_imagereturn_face_features_image_2original_image_bgrface_features_image id_cond_vit id_vit_hiddenid_cond_vit_normid_condrTs @r!process_face_embeddingsrksL UC$5$56I "I 9~9*qr  &k2U# Y'&&&= 003!!# = & &'1,566,,Q/J fg)22:>(():;>>vT"-77::t4>>qAEI #..y@UWl/mnopq !((Q(= / 484 4 9 9 ;ooe, %*[[['%.%Q"',{{2{E'J$ \\*;S=N=NO-t<FFqIEQ DII"%A!"$5$@$@BSB[B[$$79K_!2|,%W[ej"Kzz+q!T:))K)9:Kii K(bG $  r#c [U[5(a-[R"[ US9R S55n OM[R"[ R"[R"U55R S55n [U S5n U n [UUUUUUUUU U U 5 uppUR5R5nUR5nURSSS5nUR!5S-nUR#[R$5n[ R"[R"U55n XX4$)a0 Process face embeddings from an input image for inference, including alignment, feature extraction, and embedding concatenation. Args: face_helper_1: Face helper object (first helper) for alignment and landmark detection. clip_vision_model: Pre-trained CLIP vision model used for feature extraction. face_helper_2: Face helper object (second helper) for embedding extraction. eva_transform_mean: Mean values for image normalization before passing to EVA model. eva_transform_std: Standard deviation values for image normalization before passing to EVA model. app: Application instance used for face detection. device: Device (CPU or GPU) where the computations will be performed. weight_dtype: Data type of the weights for precision (e.g., `torch.float32`). img_file_path: Path to the input image file (string) or a numpy array representing an image. is_align_face: Boolean flag indicating whether face alignment should be performed (default: True). Returns: Tuple: - id_cond: Concatenated tensor of Ante face embedding and CLIP vision embedding. - id_vit_hidden: Hidden state of the CLIP vision model, a list of tensors. - image: Processed face image after feature extraction and alignment. - face_kps: Keypoints of the face detected in the image. )rRGBir(rr)r5strnparrayr convertrexif_transposer fromarrayr"rcpudetachsqueezepermutenumpyr*uint8)rvrwrxryrzr{r|r} img_file_pathrrr~rralign_crop_face_imagertensors r!process_face_embeddings_inferrs#J-%%-8@@GH001OPXXY^_` $E4 0E?V   ?;G1# & & ( / / 1F ^^ F ^^Aq! $F \\^c !F ]]288 $F  # #EOOF$; PQ (4- 8 802Q6 '$ 7 7.014+)#  Z HUlTmO1z:!!!#f%'v+ K] ppr#)i)TT)NT)T),importlib.util importlibrrrrr-PILrrtorchvision.transformsr!torchvision.transforms.functionalrrutilsr r __name__reutil find_spec_insightface_available_consisid_eva_clip_available_facexlib_availabler insightface.appr ImportErrorr rconsisid_eva_clip.constantsrrfacexlib.parsingr&facexlib.utils.face_restoration_helperrr"r8r<rrrr@r#r!rs  4?+ H "11-@L(~~778KLTXXnn..z:$F, h ii=SS t uu3H b cc.-8 6mtE3PBqr#