Ei[VSSKrSSKJr SSKJr SSKrSSKJrJr SSKJ r SSK J r J r JrJr "SS5r\R"R$S \S \S \S \4S j5r"SS5r"SS5r"SS5r"SS\5rS\R0S\S S4SjrS\R0S\\\4S \\4Sjr\R"R>S\S \4Sj5r S\S\S\S \4S jr!S(S!\"S"\S#\S$\S%\S&\\#\"\4S \4S'jjr$g))N) OrderedDict)Optional)nnTensor) functional)complete_box_iou_lossdistance_box_iou_lossFrozenBatchNorm2dgeneralized_box_iou_lossc\\rSrSrSrS\S\SS4SjrS\\ S\ \\ \\ 44S jr S r g) BalancedPositiveNegativeSampler zX This class samples batches, ensuring that they contain a fixed proportion of positives batch_size_per_imagepositive_fractionreturnNcXlX lg)z Args: batch_size_per_image (int): number of elements to be selected per image positive_fraction (float): percentage of positive elements per batch Nrr)selfrrs `/home/mande/repo/quber/.venv/lib/python3.13/site-packages/torchvision/models/detection/_utils.py__init__(BalancedPositiveNegativeSampler.__init__s %9!!2 matched_idxsc/n/nUGHyn[R"US:5Sn[R"US:H5Sn[URUR-5n[ UR 5U5nURU- n[ UR 5U5n[R"UR 5URS9SUn [R"UR 5URS9SUn XYn Xjn [R"U[RS9n [R"U[RS9nSX'SX'URU 5 URU5 GM| X#4$)a Args: matched_idxs: list of tensors containing -1, 0 or positive values. Each tensor corresponds to a specific image. -1 values are ignored, 0 are considered as negatives and > 0 as positives. Returns: pos_idx (list[tensor]) neg_idx (list[tensor]) Returns two lists of binary masks for each image. The first list contains the positive elements that were selected, and the second list the negative example. rdeviceNdtype) torchwhereintrrminnumelrandpermr zeros_likeuint8append)rrpos_idxneg_idxmatched_idxs_per_imagepositivenegativenum_posnum_negperm1perm2pos_idx_per_imageneg_idx_per_imagepos_idx_per_image_maskneg_idx_per_image_masks r__call__(BalancedPositiveNegativeSampler.__call__sa &2 "{{#9Q#>?BH{{#9Q#>?BH$33d6L6LLMG(..*G4G//'9G(..*G4GNN8>>#3HOOLXgVENN8>>#3HOOLXgVE (  ( &+%5%56LTYT_T_%` "%*%5%56LTYT_T_%` "89 " 589 " 5 NN1 2 NN1 25'38rr) __name__ __module__ __qualname____firstlineno____doc__r"floatrlistrtupler6__static_attributes__rrr r sJ3S3U3t3. T&\. eDL$v,r?rsizerMreshape decode_single) rrzr{rrrs concat_boxesbox_sumval pred_boxess rdecodeBoxCoder.decodes utUm , @   y%,, / C /44e66!9e4yyA. "C NG# Q;!))'26I'' @ Q;#++GR;J5sC#cURUR5nUSS2S4USS2S4- nUSS2S4USS2S4- nUSS2S4SU--nUSS2S4SU--nURupxpUSS2SSS24U- n USS2SSS24U- n USS2SSS24U - n USS2SSS24U - n[R"XR S9n [R"XR S9nXSS2S4-USS2S4-nXSS2S4-USS2S4-n[R "U 5USS2S4-n[R "U5USS2S4-n[R"SURURS 9U-n[R"SURURS 9U-nUU- nUU- nUU-nUU-n[R"UUUU4SS 9RS5nU$) z From a set of original boxes and encoded relative box offsets, get the decoded boxes. Args: rel_codes (Tensor): encoded boxes boxes (Tensor): reference boxes. NrFrrGrrHr~)maxrwrI) torrDr clamprkexptensorrstackflatten)rrzr{widthsheightsctr_xctr_yrNrOrPrQdxdydwdh pred_ctr_x pred_ctr_ypred_wpred_hc_to_c_hc_to_c_w pred_boxes1 pred_boxes2 pred_boxes3 pred_boxes4rs rrBoxCoder.decode_singles)q!tuQT{*1+ad +ad cFl*ad cGm+ q!$Q$w " $ q!$Q$w " $ q!$Q$w " $ q!$Q$w " $[[!5!5 6 [[!5!5 6D/)E!T'N: !T'**U1d7^; 2402D!11<<:+;+;FMMRU[[<<:+;+;FMMRU[[ 8+  8+  8+  8+ [[+{K!U[\]eefgh r)rkrD)r8r9r:r;r<mathrLr?r=rr>rrtrprrr@rArrririzs TXS[S[\gSh /UE5%78 /KP /  /1d6l1tF|1PTU[P\1V6 tF|*)v)f))rricZ\rSrSrSrSS\SS4SjjrS\S\S\4S jrS \S \S\4S jr S r g)BoxLinearCoderz The linear box-to-box transform defined in FCOS. The transformation is parameterized by the distance from the center of (square) src box to 4 edges of the target box. normalize_by_sizerNcXlg)zY Args: normalize_by_size (bool): normalize deltas by the size of src (anchor) boxes. Nr)rrs rrBoxLinearCoder.__init__s "3rrBrCc,SUSUS--nSUSUS--nX2S- nXBS- nUSU- nUSU- n[R"XVXx4SS9n UR(a1USUS- n USUS- n [R"XX4SS9n X- n U $)a Encode a set of proposals with respect to some reference boxes Args: reference_boxes (Tensor): reference boxes proposals (Tensor): boxes to be encoded Returns: Tensor: the encoded relative box offsets that can be used to decode the boxes. rH.r.rF.r.rGr}rI)r rr) rrBrCreference_boxes_ctr_xreference_boxes_ctr_ytarget_ltarget_ttarget_rtarget_brfreference_boxes_wreference_boxes_hreference_boxes_sizes rrtBoxLinearCoder.encodes!$v'>QWAX'X Y #v'>QWAX'X Y)V+<<(V+<<V$'<<V$'<<++x8FBO  ! ! / 7/&:Q Q  / 7/&:Q Q #(;;"7H\bd$ 4Grrzr{cZURURS9nSUSUS--nSUSUS--nUR(a1USUS- nUSUS- n[R"XVXV4SS9nX-nX1S- nXAS- n X1S-n XAS-n [R"XX4SS9n U $) a From a set of original boxes and encoded relative box offsets, get the decoded boxes. Args: rel_codes (Tensor): encoded boxes boxes (Tensor): reference boxes. Returns: Tensor: the predicted boxes with the encoded relative box offsets. .. note:: This method assumes that ``rel_codes`` and ``boxes`` have same size for 0th dimension. i.e. ``len(rel_codes) == len(boxes)``. rrHrrrrr}rI)rrrr r) rrzr{rrboxes_wboxes_h list_box_sizerrrrrs rrBoxLinearCoder.decodes"y/uV}uV}45uV}uV}45  ! !FmeFm3GFmeFm3G!KK7(LRTUM!1I// // // // [[+K!U[]^ rr)T) r8r9r:r;r<boolrrrtrr@rArrrrsO 3$3$3!f!!F!F##v#&#rrc t\rSrSrSrSrSr\\S.rSS\ S\ S\ S S 4S jjr S \ S \ 4S jr S\ S\ S \ S S 4SjrSrg )Matcheri9aa This class assigns to each predicted "element" (e.g., a box) a ground-truth element. Each predicted element will have exactly zero or one matches; each ground-truth element may be assigned to zero or more predicted elements. Matching is based on the MxN match_quality_matrix, that characterizes how well each (ground-truth, predicted)-pair match. For example, if the elements are boxes, the matrix may contain box IoU overlap values. The matcher returns a tensor of size N containing the index of the ground-truth element m that matches to prediction n. If there is no match, a negative value is returned. r})BELOW_LOW_THRESHOLDBETWEEN_THRESHOLDShigh_threshold low_thresholdallow_low_quality_matchesrNcvSUlSUl[R"X!:*S5 XlX lX0lg)aO Args: high_threshold (float): quality values greater than or equal to this value are candidate matches. low_threshold (float): a lower quality threshold used to stratify matches into three levels: 1) matches >= high_threshold 2) BETWEEN_THRESHOLDS matches in [low_threshold, high_threshold) 3) BELOW_LOW_THRESHOLD matches in [0, low_threshold) allow_low_quality_matches (bool): if True, produce additional matches for predictions that have only low-quality match candidates. See set_low_quality_matches_ for more details. r}rz)low_threshold should be <= high_thresholdN)rrr rrrr)rrrrs rrMatcher.__init__Ps8$& "$ m57bc,*)B&rmatch_quality_matrixcUR5S:Xa)URSS:Xa [S5e[S5eURSS9up#UR(aUR 5nOSnX R :nX R :X R:-nURX5'URX6'UR(a.Uc[R"SS5 U$URX4U5 U$)aI Args: match_quality_matrix (Tensor[float]): an MxN tensor, containing the pairwise quality between M ground-truth elements and N predicted elements. Returns: matches (Tensor[int64]): an N tensor where N[i] is a matched gt in [0, M - 1] or a negative value indicating that prediction i could not be matched. rzENo ground-truth boxes available for one of the images during trainingzANo proposal boxes available for one of the images during trainingrINFzall_matches should not be None) r$shape ValueErrorrrclonerrrrr rset_low_quality_matches_)rr matched_valsmatches all_matchesbelow_low_thresholdbetween_thresholdss rr6Matcher.__call__es % % '1 ,#))!,1 !hii !dee!5 8 8Q 8 ?  ) )!--/KK+-?-??*.@.@@\TgTgEgh'+'?'?$&*&=&=#  ) )" e%EF--gDXYrrrczURSS9upE[R"X4SS2S4:H5nUSnX'X'g)aH Produce additional matches for predictions that have only low-quality matches. Specifically, for each ground-truth find the set of predictions that have maximum overlap with it (including ties); for each prediction in that set, if it is unmatched, then match it to the ground-truth with which it has the highest quality value. rrIN)rr r!)rrrrhighest_quality_foreach_gt_ gt_pred_pairs_of_highest_qualitypred_inds_to_updates rr Matcher.set_low_quality_matches_sR)=(@(@Q(@(G%"+0;;7KjkmqjqOr7r+s(?qA'2'G$r)rrrrr)F)r8r9r:r;r<rrr"__annotations__r=rrrr6rr@rArrrr9s  #!O CuCUC_cCptC*&V&&PHHVHciHnrHrrcL^\rSrSrS\SS4U4SjjrS\S\4U4SjjrSrU=r $) SSDMatcheri thresholdrNc">[TU]XSS9 g)NF)r)superr)rr __class__s rrSSDMatcher.__init__s Orrc>[TU]U5nURSS9up4[R"UR S5[R URS9X$'U$)NrrIrrw)rr6rr arangerint64r)rrrrhighest_quality_pred_foreach_gtrs rr6SSDMatcher.__call__sa'"#78.B-E-E!-E-L*38<< + 0 0 35;;OnOuOu4 0rrA) r8r9r:r;r=rrr6r@ __classcell__)rs@rrrs1P%PDP V   rrmodelepsclUR5H n[U[5(dMXlM" g)a This method overwrites the default eps values of all the FrozenBatchNorm2d layers of the model with the provided value. This is necessary to address the BC-breaking change introduced by the bug-fix at pytorch/vision#2933. The overwrite is applied only when the pretrained weights are loaded to maintain compatibility with previous versions. Args: model (nn.Module): The model on which we perform the overwrite. eps (float): The new value of eps. N)modulesrr r)rrmodules r overwrite_epsrs'--/ f/ 0 0J"rrcURnUR5 [R"5 [ UR 55R n[R"SSUSUS4US9nU"U5n[U[R5(a[SU4/5nUR5Vs/sHofRS5PM nnSSS5 U(aUR5 W$s snf!,(df  N,=f)ak This method retrieves the number of output channels of a specific model. Args: model (nn.Module): The model for which we estimate the out_channels. It should return a single Tensor or an OrderedDict[Tensor]. size (Tuple[int, int]): The size (wxh) of the input. Returns: out_channels (List[int]): A list of the output channels of the model. rrGrr0N)trainingevalr no_gradnext parametersrzerosrrrvaluesrtrain)rr in_trainingrtmp_imgfeaturesx out_channelss rretrieve_out_channelsrs..K JJL e&&()00++q!T!Wd1g6vF> h - -"S(O#45H+3??+<=+ sB C><C9C>9C>> D vcU$NrA)rs r_fake_cast_onnxrs Hrinput orig_kvalaxisc x[RR5(d[XR U55$[R "U5UR S5n[R"[R"[R"U/URS9U4S55n[U5$)aA ONNX spec requires the k-value to be less than or equal to the number of inputs along provided dim. Certain models use the number of elements along a particular axis instead of K if K exceeds the number of elements along that axis. Previously, python's min() function was used to determine whether to use the provided k-value or the specified dim axis value. However, in cases where the model is being exported in tracing mode, python min() is static causing the model to be traced incorrectly and eventually fail at the topk node. In order to avoid this situation, in tracing mode, torch.min() is used instead. Args: input (Tensor): The original input tensor. orig_kval (int): The provided k-value. axis(int): Axis along which we retrieve the input size. Returns: min_kval (int): Appropriately selected k-value. rr) r jit is_tracingr#r_shape_as_tensorrKrMrrr)rrr axis_dim_valmin_kvals r _topk_minrs& 99   ! !9jj.//))%06@@CLyyELL)LDVDV$WYe#fhijkH 8 $$rtype box_coderanchors_per_imagematched_gt_boxes_per_imagebbox_regression_per_imagecnfc[R"US;SU35 US:Xa&URX25n[R"XFSS9$US:Xa7URX25nUb SU;aUSOSn[R "XFSUS 9$UR XB5nUb S U;aUS OS n US :Xa [XSU S 9$US:Xa [XSU S 9$[XSU S 9$)N)l1 smooth_l1cioudiougiouzUnsupported loss: rsum) reductionrbetag?)rrrgHz>r)rrr) r rrpFl1_losssmooth_l1_lossrrr r ) rrrrrrtarget_regressionrbbox_per_imagers r _box_lossr$s MM$EEI[\`[aGbc t|%334Nbyy2QVWW  %334Nb!o&C-s6{S 9X]dhii"001J^O c%j$ 6>(_djmn n 6>(_djmn n'^cilmmrr)%r collectionsrtypingrr rrtorch.nnrrtorchvision.opsrr r r r r _script_if_tracingrgrirrrModuler=rr?r"r>runusedrrstrdictr$rArrr.s # $uu< < ~,&,V,f,QW,,^ffRSSlgHgHT   4$%S/d3i< v #  %V%%3%3%@'+ n nnn!' n & n $sEz" # n nr