Ei8SSKrSSKrSSKrSSKrSSKJr SSKJr SSKJ r J r J r J r SSK rSSKrSSKJrJrJrJrJr /SQr\R0"5S;S\ \R2\\R24S\S \S \S \ \\\4S \S \S\R24Sjj5r"SS\R*5r S\RBS\"S\R&4Sjr#\R0"5SS\R2S$\R2S%\S\ \ \\ \"\\\\44\"\\\\44S\R24 S&jj5r'\R0"5S?S\R2S'\R2S(\ \\\\4S\ \ \"\\\\44S)\S\S*\ \R2S\R24S+jj5r(\R0"5S,\R2S\R24S-j5r)\R0"5S.\R2S\R24S/j5r*S\R24S0jr+S1\4S2jr,\RZS3.S\ S\"\\\\4\\ \"\\\\444S1\S4\R\S\\\\\44S5jjr/S\ SS4S6jr0S7\ S8\S\\ S944S:jr1g)@N)repeat) FunctionType)AnyBinaryIOOptionalUnion) __version__Image ImageColor ImageDraw ImageFont)_Image_fromarray make_grid save_imagedraw_bounding_boxesdraw_segmentation_masksdraw_keypoints flow_to_imagetensornrowpadding normalize value_range scale_each pad_valuereturnc4^[RR5(d2[RR5(d[ [ 5 [R "U5(dh[U[5(a<UH5n[R "U5(aM [S[U535e O[S[U535e[U[5(a[R"USS9nUR5S:XaURS5nUR5S:Xa?URS5S:Xa[R"XU4S5nURS5nUR5S:Xa.URS5S:Xa[R"XU4S5nUS La[UR!5nUb [U["5(d [S 5eS mU4S jnUS LaUH nU"Xt5 M OU"X5 [U[R$5(d [S 5eURS5S:XaUR'S5$URS5n [)X5n [+[,R."[1U 5U - 55n [+URS5U-5[+URS5U-5pURS5nUR3XU -U-X-U-4U5nSn[5U 5Hin[5U 5HWnUU :a MUR7SUU -U-X- 5R7SUU -U-X- 5R9UU5 US-nMY Mk U$)a Make a grid of images. Args: tensor (Tensor or list): 4D mini-batch Tensor of shape (B x C x H x W) or a list of images all of the same size. nrow (int, optional): Number of images displayed in each row of the grid. The final grid size is ``(B / nrow, nrow)``. Default: ``8``. padding (int, optional): amount of padding. Default: ``2``. normalize (bool, optional): If True, shift the image to the range (0, 1), by the min and max values specified by ``value_range``. Default: ``False``. value_range (tuple, optional): tuple (min, max) where min and max are numbers, then these numbers are used to normalize the image. By default, min and max are computed from the tensor. scale_each (bool, optional): If ``True``, scale each image in the batch of images separately rather than the (min, max) over all images. Default: ``False``. pad_value (float, optional): Value for the padded pixels. Default: ``0``. Returns: grid (Tensor): the tensor containing grid of images. z:tensor or list of tensors expected, got a list containing z(tensor or list of tensors expected, got rdimTzNvalue_range has to be a tuple (min, max) if specified. min and max are numbersczURXS9 URU5R[X!- S55 g)N)minmaxgh㈵>)clamp_sub_div_r&)imglowhighs N/home/mande/repo/quber/.venv/lib/python3.13/site-packages/torchvision/utils.pynorm_ipmake_grid..norm_ipTs/ JJ3J ) HHSM  s4:t4 5c>UbT"XSUS5 gT"U[UR55[UR555 g)Nrr")floatr%r&)trr.s r- norm_rangemake_grid..norm_rangeXs;&q>;q>:5>5>:r0z%tensor should be of type torch.Tensor)torchjit is_scripting is_tracing_log_api_usage_oncer is_tensor isinstancelist TypeErrortypestackr unsqueezesizecatclonetupleTensorsqueezer%intmathceilr2new_fullrangenarrowcopy_)rrrrrrrr3r4nmapsxmapsymapsheightwidth num_channelsgridkyxr.s @r-rrs> 99 ! ! # #EII,@,@,B,BI& ??6 " " fd # #q))#&`aefgah`i$jkkFtF|nUV V&$V+ zz|q!!!$ zz|q ;;q>Q YY7;F!!!$ zz|qV[[^q0FF3Q7D  ":k5+I+Ilm m 6 ;  1* v + fell + +?@@ {{1~~~a   KKNE  E  %,./ 0E A013v{{1~7O3PE;;q>L ??L5.7*BEMT[D[\^g hD A 5\uAEz KK1v:/1A B I I1u9w& eF1IAA Kr0c,\rSrSrSrSSjrSSjrSrg) _ImageDrawTV~z[ A wrapper around PIL.ImageDraw to add functionalities for drawing rotated bounding boxes. Nc $URUSUS4USUS44XCS9 [S[U5S5HCnURXXS-4XS-[U5-XS-[U5-44UUS9 ME UR XSSS9 g)Nrr"r r!)rSfillr]outlinerS) dashed_linerLlenlinepolygon)selfxyr]r_rSis r-oriented_rectangle_ImageDrawTV.oriented_rectangles 2a5"Q%.2a5"Q%.9Uq#b'1%A II%E#ba%3r7):&;RQ#b'@Q=R%ST  & RD :r0cfSn[S[U55H'nXS- upXupX{U - S-X- S--S-- nM) Sn Sn[S[U55HnXS- upXupX- S-X- S--S-nUS:dM,[X(aUOU5nX- nX- n[R"UU5nU [R "U5U--nU [R "U5U--nU(aURX4UU4/X#U5 U U- n UU-nUUpU(+nUS:aMM g)Nrr"r ?T)rLrar%rIatan2cossinrb)rdrer]rSjoint dash_length space_length total_lengthrfx1y1x2y2current_position current_dashsegment_lengthdash_length_to_drawdxdyangleend_xend_ys r-r`_ImageDrawTV.dashed_lines[ q#b'"AAYFBUFB 2g!^rw1n<D DL#  q#b'"AAYFBUFB!w1nA~=#EN 1$&).+[g&h#WW 2r*TXXe_/BBBTXXe_/BBBIIx%8$uM $77 "55B#// !!1$ #r0)NNr")NrNr)__name__ __module__ __qualname____firstlineno____doc__rgr`__static_attributes__rr0r-rZrZ~s;!0r0rZobjmodec [S[R"S555nUbUS:aURnUSn[ U5nUS:XaSOUSUS4nUR SS5nUnUbO[US 5(aUR5nO-[US 5(aUR5nO [S 5e[R"XUS USS5$[R"X5$![a SnNf=f) z A wrapper around PIL.Image.fromarray to mitigate the deprecation of the mode paramter. See: https://pillow.readthedocs.io/en/stable/releasenotes/11.3.0.html#image-fromarray-mode-parameter c38# UHn[U5v M g7fN)rH).0rXs r- #_Image_fromarray..sP/O!s1vv/Os.N) r!shaper"rstridestobytestostringz,Unable to convert obj into contiguous formatraw)rEPILLOW_VERSION_STRINGsplit Exception__array_interface__ragethasattrrr ValueErrorr frombuffer fromarray) rrPILLOW_VERSIONarrrndimrBrcontiguous_objs r-rrsP/D/J/J3/OPP!n&?N%%G 5zAIq58U1X5'')T*36  sI&&!$j))!$ !OPPNE4ANNs)){ s&C66 DDfpformatc [RR5(d2[RR5(d[ [ 5 [ U40UD6nURS5RS5RSS5RSSS5RS[R5R5n[R"U5nUR!XS9 g) a Save a given Tensor into an image file. Args: tensor (Tensor or list): Image to be saved. If given a mini-batch tensor, saves the tensor as a grid of images by calling ``make_grid``. fp (string or file object): A filename or a file object format(Optional): If omitted, the format to use is determined from the filename extension. If a file object was used instead of a filename, this parameter should always be used. **kwargs: Other arguments are documented in ``make_grid``. rjrr"r cpu)rN)r6r7r8r9r:rrmuladd_r'permutetouint8numpyr rsave)rrrkwargsrUndarrims r-rrs& 99 ! ! # #EII,@,@,B,BJ' V &v &D HHSM  s # * *1c 2 : :1a C F Fuekk Z ` ` bE  BGGBGr0imageboxeslabelscolorsr]rSfont font_size label_colorslabel_background_colors fill_labelsc SSKJs Js Jn [R R 5(d2[R R5(d[[5 [U[R5(d[S[U535eUR[R:Xd-UR!5(d[#SUR35eUR%5S:wa [#S5eUR'S5S;a [#S5eUR(S S :XaWUSS2S4USS2S 4:R+5(d&USS2S 4USS2S4:R+5(a [#S 5eUR(Sn U S:Xa[,R."S5 U$UcS/U -nO*[1U5U :wa[#SU S[1U5S35e[3Xr"r!z+Only grayscale and RGB images are supportedr#r r"zdBoxes need to be in (xmin, ymin, xmax, ymax) format. Use torchvision.ops.box_convert to convert themz/boxes doesn't contain any box. No box was drawnzNumber of boxes (z) and labels (z/) mismatch. Please specify labels for each box.) num_objectsblackz=Argument 'font_size' will be ignored since 'font' is not set. )rrB)r!r"r"TdtypescaleRGBA)d)rSr_r])r)r])r]r)+$torchvision.transforms.v2.functional transformsv2 functionalr6r7r8r9r:rr<rFr>r?rris_floating_pointrrrBranywarningswarnra _parse_colorscopyr load_defaulttruetypetileto_dtype to_pil_imagerint64tolistrZziprg rectangletextbboxtext pil_to_tensor) rrrrr]rSrrrrrF num_boxestxt_fontoriginal_dtype img_to_draw img_boxesdrawbboxcolorlabel label_colorlabel_bg_color draw_method fill_color box_marginmarginlefttoprightbottomouts r-rrs`54 99 ! ! # #EII,@,@,B,B/0 eU\\ * */U }=>>kkU[[(E,C,C,E,EG }UVV  >?? Af $FGG RA E!Q$K%1+$=#B#B#D#DqRStW\]^`a]aWbIbHgHgHiHi r   AIA~ GH  ~04v /A V ! {.V Et u  6 9F{$\\w\ef {{} ."/0G"_"(++- |  MMY Z))+%%4o2F zz!} 5),[[N'' 5 4 @..'K%,,.I K0K(;>y&Zfr?rrrrrrBrboolrsumrrrrrdetachrDrr) rrrr num_masksoverlapping_masksrrrmaskrs r-rrsB2 99 ! ! # #EII,@,@,B,B34 eU\\ * *:4;-HIIkkU[[(E,C,C,E,EG }UVV  >?? aA STT zzQdAqj! zzQNOO {{ejj @ NOO {{235;;rs++VWW QI a (1,A~ IJ [[N#6WWE  UFW  ,,.&&(K5) $QW~ AtG*)*K$$% 1u9  3 3C 66. !!s )J keypoints connectivityradius visibilityc [RR5(d2[RR5(d[ [ 5 [ U[R5(d[S[U535eUR[R:Xd-UR5(d[SUR35eUR5S:wa [S5eUR5SS:wa [S5eUR S:wa [S5eUc0[R""UR$SS [R&S 9nUR S:XaUR)S 5nUR S :wa[S UR 35eUR$UR$SS :wa'[S UR$<SUR$<35eURnUR(aSSKJn U"U[RSS9nUR/SS S5R15R35n [4R6"U 5n [8R:"U 5n UR=[R>5RA5n UR15R'5RA5n [CX5Hup[CX5HEunnU(dMUSU- nUSU-nUSU- nUSU-nU REUUUU/USSS9 MG U(dMbUHanUUS(a UUS(dMUUSSnUUSSnUUSSnUUSSnU RGUU4UU44US9 Mc M [RH"[JRL"U 55R/S SS5nUR(aW"UUSS9nU$)a Draws Keypoints on given RGB image. The image values should be uint8 in [0, 255] or float in [0, 1]. Keypoints can be drawn for multiple instances at a time. This method allows that keypoints and their connectivity are drawn based on the visibility of this keypoint. Args: image (Tensor): Tensor of shape (3, H, W) and dtype uint8 or float. keypoints (Tensor): Tensor of shape (num_instances, K, 2) the K keypoint locations for each of the N instances, in the format [x, y]. connectivity (List[Tuple[int, int]]]): A List of tuple where each tuple contains a pair of keypoints to be connected. If at least one of the two connected keypoints has a ``visibility`` of False, this specific connection is not drawn. Exclusions due to invisibility are computed per-instance. colors (str, Tuple): The color can be represented as PIL strings e.g. "red" or "#FF00FF", or as RGB tuples e.g. ``(240, 10, 157)``. radius (int): Integer denoting radius of keypoint. width (int): Integer denoting width of line connecting keypoints. visibility (Tensor): Tensor of shape (num_instances, K) specifying the visibility of the K keypoints for each of the N instances. True means that the respective keypoint is visible and should be drawn. False means invisible, so neither the point nor possible connections containing it are drawn. The input tensor will be cast to bool. Default ``None`` means that all the keypoints are visible. For more details, see :ref:`draw_keypoints_with_visibility`. Returns: img (Tensor[C, H, W]): Image Tensor with keypoints drawn. rrr!rrrz0keypoints must be of shape (num_instances, K, 2)Nrrr z9visibility must be of shape (num_instances, K). Got ndim=zikeypoints and visibility must have the same dimensionality for num_instances and K. Got visibility.shape=z and keypoints.shape=)rTrr"r^)rS)'r6r7r8r9r:rr<rFr>r?rrrrrrBronesrrrGrrrrrr rr Drawrrrrellipserb from_numpynparray)rrrrrrSrrrrrrimg_kptsimg_viskpt_instvis_inst kpt_coordkp_visrrrtrsru connection start_pt_x start_pt_yend_pt_xend_pt_yrs r-rrsT 99 ! ! # #EII,@,@,B,BN+ eU\\ * *:4;-HIIkkU[[(E,C,C,E,EG }UVV  >?? aA STT~~KLLZZ  4EJJG ! ''+ !TU_UdUdTefgg9??3B// %##%%;9??*< >  [[N''Aekk> MM!Q " & & ( . . 0E//%(K >>+ &D||EKK(//1Hnn##%,,.G!(4!$X!8 Iv1&B1&B1&B1&B LL"b"b)AL N"9 <*  A/*Q-9P%jm4Q7 %jm4Q7 #JqM215#JqM215  *-(/CD+52   288K0 1 9 9!Q BC''s.= Jr0flowc"UR[R:wa[SURS35eURnUR S:XaUSnUR S:wdURSS:wa[SUS35e[R "US-SS 9R5R5n[R"UR5RnXU-- n[U5n[U5S:XaUS nU$) a. Converts a flow to an RGB image. Args: flow (Tensor): Flow of shape (N, 2, H, W) or (2, H, W) and dtype torch.float. Returns: img (Tensor): Image Tensor of dtype uint8 where each color corresponds to a given flow direction. Shape is (N, 3, H, W) or (3, H, W) depending on the input. z)Flow should be of dtype torch.float, got rr!Nr#r"r zURupp4URn[R"USX44[RUS9n[ 5R U5nURSn[R"US-SS9R5n [R"USS2SSS2SS24*USS2SSS2SS24*5[R- n U S-S- US- -n [R"U 5R [R5n U S-n SXU:H'X- n[URS5H[nUSS2U4nUU S- nUU S- nSU- U-UU--nSU SU- -- n[R"S U-5USS2USS2SS24'M] U$) z Converts a batch of normalized flow to an RGB image. Args: normalized_flow (torch.Tensor): Normalized flow tensor of shape (N, 2, H, W) Returns: img (Tensor(N, 3, H, W)): Flow visualization image of dtype uint8. r!rrr r"rNgo@r)rrr6zerosr_make_colorwheelrrrrkpifloorlongrL)r!N_HWr flow_image colorwheelnum_colsnormafkk0k1fctmpcol0col1cols r-rr|s!&&JA!  # #FaA\VLJ!#&&v.J"H 99_a'Q / 4 4 6D _Q1aZ00?1aA:3N2NORWRZRZZA a%11 %B R  EJJ 'B aBBX~ A :##A& 'A2w2w1unq4x'$!c'""!&S3Y!7 1aA: ( r0cSnSnSnSnSnSnX-U-U-U-U-n[R"US45nSnSUSU2S4'[R"S[R"SU5-U- 5USU2S 4'X-nS[R"S[R"SU5-U- 5- XxX-2S4'SXxX-2S 4'X-nSXxX-2S 4'[R"S[R"SU5-U- 5XxX-2S 4'X-nS[R"S[R"U5-U- 5- XxX-2S 4'SXxX-2S 4'X-nSXxX-2S 4'[R"S[R"SU5-U- 5XxX-2S4'X-nS[R"S[R"U5-U- 5- XxX-2S 4'SXxX-2S4'U$) a Generates a color wheel for optical flow visualization as presented in: Baker et al. "A Database and Evaluation Methodology for Optical Flow" (ICCV, 2007) URL: http://vision.middlebury.edu/flow/flowEval-iccv07.pdf. Returns: colorwheel (Tensor[55, 3]): Colorwheel Tensor. r#r r!rrr"r )r6r#r&arange) RYYGGCCBBMMRncolsr-r9s r-r$r$s  B B B B B B GbL2  "R 'EeQZ(J CJqtQw++cELLB,?&?"&DEJqtQw (C$'%++cELLB.s/ACs) rHr<r=rarrEstrr getrgbr)rrrrs r-rrs,6~(5 FD ! ! v; $_`cdj`k_llop{o||}~  % - -Z[aZbbcdee FE " "s6{a'7XY_X``abccK'Y_ `Y_PU*UC*@*@j&eKY_F ` =CDVE%///VD MaDs 77D%D*cURnURS5(dSU3nURRn[ U[ 5(a URn[ RRUSU35 g)a Logs API usage(module and name) within an organization. In a large ecosystem, it's often useful to track the PyTorch and TorchVision APIs usage. This API provides the similar functionality to the logging module in the Python stdlib. It can be used for debugging purpose to log which methods are used and by default it is inactive, unless the user manually subscribes a logger via the `SetAPIUsageLogger method `_. Please note it is triggered only once for the same API call within a process. It does not collect any data from open-source users since it is no-op by default. For more information, please refer to * PyTorch note: https://pytorch.org/docs/stable/notes/large_scale_deployments.html#api-usage-logging; * Logging policy: https://github.com/pytorch/vision/issues/5052; Args: obj (class instance or method): an object to extract info from. torchvisionztorchvision.internal.rN) r startswith __class__rr<rr6_Cr:)rmodulenames r-r:r:sl"^^F   ] + +(1 == ! !D#|$$|| HH  F81TF!34r0rXn.c[U[RR5(a [ U5$[ [ X55$)a7 Make n-tuple from input x. If x is an iterable, then we just convert it to tuple. Otherwise, we will make a tuple of length n, all with value of x. reference: https://github.com/pytorch/pytorch/blob/master/torch/nn/modules/utils.py#L8 Args: x (Any): input value n (int): length of the resulting tuple )r< collectionsabcIterablerEr)rXrVs r- _make_ntupler[s3![__--..Qx  r0)r FNFgr) NNFr"NNNNF)g?N)NNr r!N)2rXrIpathlibr itertoolsrtypesrtypingrrrrrr r6PILr rr r r r __all__no_gradrFr=rHrrEr2rrZndarrayrMrPathrrrrrrr$rHrrrr:r[rr0r-rfs 11 ]] -1b %,,U\\ 22 3b bb b %S/* b  bb \\bbJ009&&00fL* L* L* [[L*^! %,,U\\ 22 3 c7<<)* SM  6#'ae #gkrvA <<A <<A T#Y A U4c5c3+?&? @A3cSVX[mH\\] ^ A 4. A  A 3-A}A5eCsC}1E,E&F!GeTWY\^aTaNb!bcdA&eDsE#sC-W8X3YY Z,, ;; ,  %S#  ,^5S5T54 C C E#s(O r0