hG SSKrSSKrSSKJr SSKJr SSKJr SSKJrJ r J r J r J r J r JrJr SSKJr S \S \S \S \4S jrS \S \S\S \S \4 Sjr\R*R,S \S \S\S \S \4 Sj5r\R*R,S \S \S\S \S \4 Sj5rS \S\S \4SjrS \S\\\4S \4SjrS \S\S\S \4SjrS \S \4SjrS\S\S \\\44Sjr S\S\S \4Sjr!S\S\S \4Sjr"S%S\S\S\S \4S jjr#S%S\S\S\S \4S!jjr$S%S\S\S\S \\\44S"jjr%S#\RS \R4S$jr&g)&N)Tensor)_assert_has_ops)_log_api_usage_once)_box_cxcywh_to_xyxy_box_cxcywhr_to_xywhr_box_xywh_to_xyxy_box_xywhr_to_cxcywhr_box_xywhr_to_xyxyxyxy_box_xyxy_to_cxcywh_box_xyxy_to_xywh_box_xyxyxyxy_to_xywhr)_upcastboxesscores iou_thresholdreturnc[RR5(d2[RR5(d[ [ 5 [ 5 [RRR XU5$)a Performs non-maximum suppression (NMS) on the boxes according to their intersection-over-union (IoU). NMS iteratively removes lower scoring boxes which have an IoU greater than ``iou_threshold`` with another (higher scoring) box. If multiple boxes have the exact same score and satisfy the IoU criterion with respect to a reference box, the selected box is not guaranteed to be the same between CPU and GPU. This is similar to the behavior of argsort in PyTorch when repeated values are present. Args: boxes (Tensor[N, 4])): boxes to perform NMS on. They are expected to be in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. scores (Tensor[N]): scores for each one of the boxes iou_threshold (float): discards all overlapping boxes with IoU > iou_threshold Returns: Tensor: int64 tensor with the indices of the elements that have been kept by NMS, sorted in decreasing order of scores ) torchjit is_scripting is_tracingrnmsrops torchvision)rrrs M/var/www/auris/envauris/lib/python3.13/site-packages/torchvision/ops/boxes.pyrrsT2 99 ! ! # #EII,@,@,B,BC  99 $ $UM BBidxscp[RR5(d2[RR5(d[ [ 5 UR 5URRS:XaSOS:a&[R"5(d [XX#5$[XX#5$)a Performs non-maximum suppression in a batched fashion. Each index value correspond to a category, and NMS will not be applied between elements of different categories. Args: boxes (Tensor[N, 4]): boxes where NMS will be performed. They are expected to be in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. scores (Tensor[N]): scores for each one of the boxes idxs (Tensor[N]): indices of the categories for each one of the boxes. iou_threshold (float): discards all overlapping boxes with IoU > iou_threshold Returns: Tensor: int64 tensor with the indices of the elements that have been kept by NMS, sorted in decreasing order of scores cpuii) rrrrr batched_nmsnumeldevicetyper _is_tracing_batched_nms_vanilla_batched_nms_coordinate_trick)rrrrs rr"r"3s}0 99 ! ! # #EII,@,@,B,BK( {{} 1 1U :HQ\QhQhQjQj#E4GG,UDPPrcLUR5S:Xa.[R"S[RURS9$UR 5nUR U5U[R"S5R U5--nXSS2S4-n[XaU5nU$)Nr)rdtyper$r) r#remptyint64r$maxtotensorr)rrrrmax_coordinateoffsets boxes_for_nmskeeps rr(r(Vs {{}{{4u{{5<<HHYY[Nggena1C1CE1J JKGAtG,,M }m 4D KrcH[R"U[RS9n[R"U5H6n[R"X%:H5Sn[ XXU5nSXFU'M8 [R"U5SnXUR SS9S$)N)r+rT) descendingr)r zeros_likebooluniquewherersort) rrrr keep_maskclass_id curr_indicescurr_keep_indices keep_indicess rr'r'js  uzz:ILL&{{4#34Q7  3V5I=Y59 012';;y)!,L |,11T1B1E FFrmin_sizec>[RR5(d2[RR5(d[ [ 5 USS2S4USS2S4- USS2S4USS2S4- p2X!:X1:-n[R "U5SnU$)a# Remove every box from ``boxes`` which contains at least one side length that is smaller than ``min_size``. .. note:: For sanitizing a :class:`~torchvision.tv_tensors.BoundingBoxes` object, consider using the transform :func:`~torchvision.transforms.v2.SanitizeBoundingBoxes` instead. Args: boxes (Tensor[N, 4]): boxes in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. min_size (float): minimum size Returns: Tensor[K]: indices of the boxes that have both sides larger than ``min_size`` Nrrr)rrrrrremove_small_boxesr:)rrAwshsr4s rrDrD{s$ 99 ! ! # #EII,@,@,B,B./ 1a4[5A; &ad eAqDk(A Nr~ .D ;;t Q D Krsizec [RR5(d2[RR5(d[ [ 5 UR 5nUSSSS24nUSSSS24nUupV[R"5(a[R"U[R"SURURS95n[R"U[R"X`RURS95n[R"U[R"SURURS95n[R"U[R"XPRURS95nO URSUS9nURSUS9n[R"X44US9nUR!UR"5$) a Clip boxes so that they lie inside an image of size ``size``. .. note:: For clipping a :class:`~torchvision.tv_tensors.BoundingBoxes` object, consider using the transform :func:`~torchvision.transforms.v2.ClampBoundingBoxes` instead. Args: boxes (Tensor[N, 4]): boxes in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. size (Tuple[height, width]): size of the image Returns: Tensor[N, 4]: clipped boxes .rNrrr*)minr.)dim)rrrrrclip_boxes_to_imagerJrr&r.r0r+r$rIclampstackreshapeshape)rrGrJboxes_xboxes_yheightwidth clipped_boxess rrKrKs] 99 ! ! # #EII,@,@,B,B/0 ))+CCAIGCAIGMF  ))GU\\!5;;u||%\]))GU\\%{{SXS_S_%`a))GU\\!5;;u||%\]))GU\\& TYT`T`%ab--A5-1--A6-2KK 2L 99 ! ! # #EII,@,@,B,BK(L!W%@QRXQYYfgnfopqq {{} A !%(< L;  #E*8 L7  !%(4 L3  #E*0 L/  !%(#E** L)  #E*!%($ L# " "%e, L " "%e, L % %%e,// <&u- L % %&u-00=%e, L # #&u- L # #&u- L"$Z[\]^[_Z``mnopqnrms"tuurc[RR5(d2[RR5(d[ [ 5 [ U5nUSS2S4USS2S4- USS2S4USS2S4- -$)aT Computes the area of a set of bounding boxes, which are specified by their (x1, y1, x2, y2) coordinates. Args: boxes (Tensor[N, 4]): boxes for which the area will be computed. They are expected to be in (x1, y1, x2, y2) format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. Returns: Tensor[N]: the area for each box NrrrCr)rrrrrbox_arear)rs rrereso 99 ! ! # #EII,@,@,B,BH% ENE !Q$K%1+ %%1+ad *C DDrboxes1boxes2c`[U5n[U5n[R"USS2SSS24USS2SS245n[R"USS2SSS24USS2SS245n[ XT- 5R SS9nUSS2SS2S4USS2SS2S4-nUSS2S4U-U- nXx4$)NrrrIr)rerr.rIrrL) rfrgarea1area2ltrbwhinterunions r_box_inter_unionrq&s V E V E 6!T2A2+&q"1"u 6B 6!T12+&q!"u 6B    A  &B q!QwK"Q1W+ %E !T'NU "U *E <rc[RR5(d2[RR5(d[ [ 5 [ X5up#X#- nU$)a Return intersection-over-union (Jaccard index) between two sets of boxes. Both sets of boxes are expected to be in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. Args: boxes1 (Tensor[N, 4]): first set of boxes boxes2 (Tensor[M, 4]): second set of boxes Returns: Tensor[N, M]: the NxM matrix containing the pairwise IoU values for every element in boxes1 and boxes2 )rrrrrbox_iourq)rfrgrorpious rrsrs5sI 99 ! ! # #EII,@,@,B,BG$#F3LE -C Jrc[RR5(d2[RR5(d[ [ 5 [ X5up#X#- n[R"USS2SSS24USS2SS245n[R"USS2SSS24USS2SS245n[Xe- 5RSS9nUSS2SS2S4USS2SS2S4-nXHU- U- - $)a Return generalized intersection-over-union (Jaccard index) between two sets of boxes. Both sets of boxes are expected to be in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. Args: boxes1 (Tensor[N, 4]): first set of boxes boxes2 (Tensor[M, 4]): second set of boxes Returns: Tensor[N, M]: the NxM matrix containing the pairwise generalized IoU values for every element in boxes1 and boxes2 Nrrrir) rrrrrgeneralized_box_iourqrIr.rrL) rfrgrorprtltirbiwhiareais rrvrvKs 99 ! ! # #EII,@,@,B,B/0#F3LE -C ))F1dBQB;'2A2 7C ))F1dAB;'12 7C #)  " "q " )C 1aL3q!Qw< 'E %-5( ((repsc[RR5(d2[RR5(d[ [ 5 [ U5n[ U5n[XU5up4USS2SS4USS2SS4- nUSS2SS4USS2SS4- nUSS2S4USS2S4- nUSS2S4USS2S4- nS[RS-- [R"[R"XV- 5[R"Xx- 5- S5-n [R"5 U SU- U -U-- n SSS5 UW U -- $!,(df  N=f)a Return complete intersection-over-union (Jaccard index) between two sets of boxes. Both sets of boxes are expected to be in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. Args: boxes1 (Tensor[N, 4]): first set of boxes boxes2 (Tensor[M, 4]): second set of boxes eps (float, optional): small number to prevent division by zero. Default: 1e-7 Returns: Tensor[N, M]: the NxM matrix containing the pairwise complete IoU values for every element in boxes1 and boxes2 NrrrCr) rrrrrcomplete_box_iour _box_diou_ioupipowatanno_grad) rfrgr{diourtw_predh_predw_gth_gtvalphas rr~r~isB 99 ! ! # #EII,@,@,B,B,- V_F V_Ffc2ID AtQJ &D!"4 4F AtQJ &D!"4 4F !Q$<&A, &D !Q$<&A, &D ehhk eii 6?(CejjQUQ\F](]_`aaA QWq[3&'  %!)  s <E E)c[RR5(d2[RR5(d[ [ 5 [ U5n[ U5n[XUS9up4U$)a Return distance intersection-over-union (Jaccard index) between two sets of boxes. Both sets of boxes are expected to be in ``(x1, y1, x2, y2)`` format with ``0 <= x1 < x2`` and ``0 <= y1 < y2``. Args: boxes1 (Tensor[N, 4]): first set of boxes boxes2 (Tensor[M, 4]): second set of boxes eps (float, optional): small number to prevent division by zero. Default: 1e-7 Returns: Tensor[N, M]: the NxM matrix containing the pairwise distance IoU values for every element in boxes1 and boxes2 )r{)rrrrrdistance_box_iourr)rfrgr{r_s rrrsW 99 ! ! # #EII,@,@,B,B,- V_F V_FF4GD Krcx[X5n[R"USS2SSS24USS2SS245n[R"USS2SSS24USS2SS245n[ XT- 5R SS9nUSS2SS2S4S-USS2SS2S4S--U-nUSS2S4USS2S4-S- nUSS2S4USS2S4-S- n USS2S4USS2S4-S- n USS2S4USS2S4-S- n [ USS2S4U SSS24- 5S-[ U SS2S4U SSS24- 5S--n X)'rrrtorchvision.extensionrutilsr _box_convertrr r r r r rr_utilsrrrr"r_script_if_tracingr(r'rDtupleintrKstrr^rerqrsrvr~rrrrrrs 1'   CvCvCeCC> Q  Q  Q  Q Q  QF      & G  G  G  G G  G G f&4!.v!.U38_!.!.HUvUsUSUVUpEFEvE* V V ffn8M FFv,)))6)<VV%6BVV%62M&M&MuMfV\nH]M&%%,,%5<<%r