h dZddlZddlZddlmZddlmZddlmZm Z m Z m Z m Z ddl ZddlZddlmZmZmZmZmZmZej.gddz ZdNd ej2d ej2d ed edej2f dZdOd ej:d ej:d edej:fdZ dPd ej:d ej:dedededed edej:fdZdOdej:dej:d edej:fdZ dOdej:dej:dej:de ed edej:f dZ!dej:de ej:ej:ej:ffdZ"dNd ej:d!ej:ded edej:f d"Z# dOd e ej:ej2fd!e ej:ej2fd edej:fd#Z$dQd ede eeffd$Z%Gd%d&eZ&dRd'ej2d(edej2fd)Z'eed*idfd+ej2d,ej2d-ej2d.ed/e e(e)ff d0Z*eed1id2d3dfd+ej2d,ej2d.ed/e e(e)fd4e)d5e)f d6Z+d7e ed8e ede eej2ej2ffd9Z,d deid:d;fdej2d?ej2d@ed.ed/e e(e)fd edAe)de fdBZ-GdCd3eZ.GdDdEeeZ/GdFdGe/Z0GdHdIe/Z1GdJdKeeZ2GdLdMe/Z3y)SzModel validation metrics.N) defaultdict)Path)AnyDictListTupleUnion)LOGGERDataExportMixin SimpleClass TryExceptchecks plt_settings)gp= ף??rffffff?rHzG?r ףp= ?rףp= ?rQ?rףp= ?r{Gz?rg$@Fbox1box2iouepsreturnc|j\}}}}|j\}} } } tj|dddf| tj|dddf|z j dtj|dddf| tj|dddf| z j dz} | |z | | z z} |r||z ||z z}| |dddfz| z } | | |zz S)aD Calculate the intersection over box2 area given box1 and box2. Args: box1 (np.ndarray): A numpy array of shape (N, 4) representing N bounding boxes in x1y1x2y2 format. box2 (np.ndarray): A numpy array of shape (M, 4) representing M bounding boxes in x1y1x2y2 format. iou (bool, optional): Calculate the standard IoU if True else return inter_area/box2_area. eps (float, optional): A small value to avoid division by zero. Returns: (np.ndarray): A numpy array of shape (N, M) representing the intersection over box2 area. Nr)Tnpminimummaximumclip)rrrrb1_x1b1_y1b1_x2b1_y2b2_x1b2_y1b2_x2b2_y2 inter_areaarea box1_areas W/var/www/html/eduruby.in/venv/lib/python3.12/site-packages/ultralytics/utils/metrics.pybbox_ioar/s"&E5%!%E5%**U1d7^U3bjjq$wQV6WW]]^_` 5D>5)BJJuQW~u,MM d1gJ EMeem ,D U]uu}5 i4((:5  $$c|jjdjdd|jjdjddc\}}\}}tj||tj ||z j djd}|||z jd||z jdz|z |zz S)aG Calculate intersection-over-union (IoU) of boxes. Args: box1 (torch.Tensor): A tensor of shape (N, 4) representing N bounding boxes in (x1, y1, x2, y2) format. box2 (torch.Tensor): A tensor of shape (M, 4) representing M bounding boxes in (x1, y1, x2, y2) format. eps (float, optional): A small value to avoid division by zero. Returns: (torch.Tensor): An NxM tensor containing the pairwise IoU values for every element in box1 and box2. References: https://github.com/pytorch/vision/blob/main/torchvision/ops/boxes.py r)float unsqueezechunktorchminmaxclamp_prod)rrra1a2b1b2inters r.box_iourA5s"//288A> @V@VWX@Y@_@_`acd@eHRhr2 YYr2 2r!2 2 : :1 = B B1 EE R"WNN1%bq(99EACG HHr0xywhGIoUDIoUCIoUc6|ru|jdd|jddc\}}} } \} } } }| dz | dz | dz |dz f\}}}}||z ||z||z ||zf\}}}}| |z | |z| |z | |zf\}}}}nH|jdd\}}}}|jdd\}}}}||z ||z |z} } ||z ||z |z}} |j||j|z jd|j||j|z jdz}| | z| |zz|z |z}||z }|s|s|rK|j||j|z }|j||j|z }|s|r|j d|j dz|z} ||z|z |z j d||z|z |z j dzdz }!|rdt j dzz | |z j| | z jz j dz}"tj5|"|"|z d|zzz }#ddd||!| z |"#zzz S||!| z z S||z|z}$||$|z |$z z S|S#1swY4xYw)a Calculate the Intersection over Union (IoU) between bounding boxes. This function supports various shapes for `box1` and `box2` as long as the last dimension is 4. For instance, you may pass tensors shaped like (4,), (N, 4), (B, N, 4), or (B, N, 1, 4). Internally, the code will split the last dimension into (x, y, w, h) if `xywh=True`, or (x1, y1, x2, y2) if `xywh=False`. Args: box1 (torch.Tensor): A tensor representing one or more bounding boxes, with the last dimension being 4. box2 (torch.Tensor): A tensor representing one or more bounding boxes, with the last dimension being 4. xywh (bool, optional): If True, input boxes are in (x, y, w, h) format. If False, input boxes are in (x1, y1, x2, y2) format. GIoU (bool, optional): If True, calculate Generalized IoU. DIoU (bool, optional): If True, calculate Distance IoU. CIoU (bool, optional): If True, calculate Complete IoU. eps (float, optional): A small value to avoid division by zero. Returns: (torch.Tensor): IoU, GIoU, DIoU, or CIoU values depending on the specified flags. r3rr2N) r6r r!r:powmathpiatanr7no_grad)%rrrBrCrDrErx1y1w1h1x2y2w2h2w1_h1_w2_h2_r#r%r$r&r'r)r(r*r@unionrcwchc2rho2valphac_areas% r.bbox_iourbMs> -1ZZ2-> 1b@Q*RR*2r2r!VR!VR!VR!V;S#s%'#XrCxc28%K"ueU%'#XrCxc28%K"ueU%)ZZ2%6"ueU%)ZZ2%6"ueU  3B  3B]]5 !EMM%$8 8 @ @ C eu}}U33 fQi E Gb2g  % +E %-C tt ]]5 !EMM%$8 8 ]]5 !EMM%$8 8 4RVVAY&,B&.33A6%%-%:ORW:W9\9\]^9__D!^b(8BG>>;K(K'P'PQR'SS]]_6SAG!45E6dRi!e)344? "b3fun... J 66s JJmask1mask2ctj||jjd}|j ddddf|j ddz|z }|||zz S)a Calculate masks IoU. Args: mask1 (torch.Tensor): A tensor of shape (N, n) where N is the number of ground truth objects and n is the product of image width and height. mask2 (torch.Tensor): A tensor of shape (M, n) where M is the number of predicted objects and n is the product of image width and height. eps (float, optional): A small value to avoid division by zero. Returns: (torch.Tensor): A tensor of shape (N, M) representing masks IoU. rr2N)r7matmulrr:sum)rcrdr intersectionrZs r.mask_iouris_<<uww/66q9L YYq\!T' "UYYq\$%7 7< GE 53; ''r0kpt1kpt2r,sigmac|ddddddf|dz jd|ddddddf|dz jdz}tj||j|j}|ddk7}|d|zjd|ddddf|zzdzz }| j |dddfzj d |j d dddf|zz S) aP Calculate Object Keypoint Similarity (OKS). Args: kpt1 (torch.Tensor): A tensor of shape (N, 17, 3) representing ground truth keypoints. kpt2 (torch.Tensor): A tensor of shape (M, 17, 3) representing predicted keypoints. area (torch.Tensor): A tensor of shape (N,) representing areas from ground truth. sigma (list): A list containing 17 values representing keypoint scales. eps (float, optional): A small value to avoid division by zero. Returns: (torch.Tensor): A tensor of shape (N, M) representing keypoint similarities. Nr).rr3r2).r2)devicedtype).r3rH)rIr7tensorrnroexprg)rjrkr,rlrdkpt_maskes r.kpt_iourus aq!m tF| +003tAtQM7JTRX\7Y6^6^_`6aaA LLt{{$** EEF|q H a%i__Q 44 #6#< = ABARHHJ!T'* * / / 3x||B7G47PSV7V WWr0boxescntj|ddddfjddz |ddddffd}|jdd\}}}|j }|j }|jd}|jd}||z||zz||z||zz||z |z|zfS)a( Generate covariance matrix from oriented bounding boxes. Args: boxes (torch.Tensor): A tensor of shape (N, 5) representing rotated bounding boxes, with xywhr format. Returns: (torch.Tensor): Covariance matrices corresponding to original rotated bounding boxes. Nr3rG rHdimr2)r7catrIsplitcossin) rvgbbsabcr}r~cos2sin2s r._get_covariance_matrixrs 99eAqsFm''*R/q!"u>B GDjjj#GAq! %%'C %%'C 771:D 771:D t8a$h D1t8 3a!es]S5H HHr0obb1obb2c|dddfjdd\}}|dddfjdd\}}t|\}} } t|\} } } || z||z jdz| | z||z jdzz|| z| | zz| | zjdz |zz dz}| | z||z z||z z|| z| | zz| | zjdz |zz dz}|| z| | zz| | zjdz d || z| jdz jd | | z| jdz jd zj z|zz |zj dz}||z|zj |d }d | jz |zj }d|z }|r|ddd fjdd\}}|ddd fjdd\}}d tjdzz ||z j||z jz jdz}tj5|||z d|zzz }ddd||zz S|S#1swYxYw) a= Calculate probabilistic IoU between oriented bounding boxes. Args: obb1 (torch.Tensor): Ground truth OBBs, shape (N, 5), format xywhr. obb2 (torch.Tensor): Predicted OBBs, shape (N, 5), format xywhr. CIoU (bool, optional): If True, calculate CIoU. eps (float, optional): Small value to avoid division by zero. Returns: (torch.Tensor): OBB similarities, shape (N,). Notes: OBB format: [center_x, center_y, width, height, rotation_angle]. References: https://arxiv.org/pdf/2106.06072v1.pdf .Nr3r2rHryr?rGrY@?) r|rrIr:sqrtlogclamprqrJrKrLr7rM)rrrErrNrOrRrSr<r>c1r=r?r]t1t2t3bdhdrrPrQrTrUr_r`s r.probiours&#rr']  +FB #rr']  +FB'-JBB'-JBB r'b2g]]1% %bR"WMM!4D(D D"r'VX[]V]I^bdgibiananopaqIqtwIwx  B GR BG ,"r'b2g1F"r'WXIY1Y\_1_ `dg gB r'b2g "r'q!1 1 b266!9$,,Q/27RVVAY3F2N2Nq2QQWWY Y\_ _ a   cec  B r'B,  c5 )B " c ! ' ' )B b&C c1Q3h%%aR%0Bc1Q3h%%aR%0B !^b 0BG>>3C CHHK K ]]_ .SAG,-E .QY J . .s "JJ cjt|tjrtj|n|}t|tjrtj|n|}|dddfj dd\}}d|dddfj ddD\}}t |\}}} dt |D\} } } || z||z jdz|| z||z jdzz|| z|| zz| | zjdz |zz d z} | | z||z z||z z|| z|| zz| | zjdz |zz d z}|| z|| zz| | zjdz d ||z| jdz jd | | z| jdz jd zjz|zz |zjd z}| |z|zj|d }d| jz |zj}d|z S)a$ Calculate the probabilistic IoU between oriented bounding boxes. Args: obb1 (torch.Tensor | np.ndarray): A tensor of shape (N, 5) representing ground truth obbs, with xywhr format. obb2 (torch.Tensor | np.ndarray): A tensor of shape (M, 5) representing predicted obbs, with xywhr format. eps (float, optional): A small value to avoid division by zero. Returns: (torch.Tensor): A tensor of shape (N, M) representing obb similarities. References: https://arxiv.org/pdf/2106.06072v1.pdf .Nr3r2rHryc3DK|]}|jddywrHNsqueeze.0xs r. z batch_probiou..s JaaiimD! J c3DK|]}|jddywrrrs r.rz batch_probiou..sL!!))B-%LrrrrGrrr) isinstancerndarrayr7 from_numpyr|rrIr:rrrrq)rrrrNrOrRrSr<r>rr=r?r]rrrrrs r. batch_probioursS"&0bjj%A5  D !tD%/bjj%A5  D !tD #rr']  +FB J4RaR=+>+>qb+>+I JFB'-JBBL/Ed/KLJBB r'b2g]]1% %bR"WMM!4D(D D"r'VX[]V]I^bdgibiananopaqIqtwIwx  B GR BG ,"r'b2g1F"r'WXIY1Y\_1_ `dg gB r'b2g "r'q!1 1 b266!9$,,Q/27RVVAY3F2N2Nq2QQWWY Y\_ _ a   cec  B r'B,  c5 )B " c ! ' ' )B r6Mr0cdd|zz d|zfS)a Compute smoothed positive and negative Binary Cross-Entropy targets. Args: eps (float, optional): The epsilon value for label smoothing. Returns: pos (float): Positive label smoothing BCE target. neg (float): Negative label smoothing BCE target. References: https://github.com/ultralytics/yolov3/issues/238#issuecomment-598028441 rr)rs r. smooth_bcer(s s?C#I %%r0c eZdZdZgddfdeeefdedefdZded eee fd ed d fd Z de e jde e jd d fdZ d"deee jfd eee fdeded d f dZdZd eej(ej(ffdZde jdeded d fdZeded#dedefdZdZd$ded ed e eeeffd!Zy )%ConfusionMatrixa  A class for calculating and updating a confusion matrix for object detection and classification tasks. Attributes: task (str): The type of task, either 'detect' or 'classify'. matrix (np.ndarray): The confusion matrix, with dimensions depending on the task. nc (int): The number of category. names (List[str]): The names of the classes, used as labels on the plot. matches (dict): Contains the indices of ground truths and predictions categorized into TP, FP and FN. detectFnamestask save_matchescB||_t||_|jdk(r+tj|j|jfn0tj|jdz|jdzf|_||_|ri|_yd|_y)aW Initialize a ConfusionMatrix instance. Args: names (Dict[int, str], optional): Names of classes, used as labels on the plot. task (str, optional): Type of task, either 'detect' or 'classify'. save_matches (bool, optional): Save the indices of GTs, TPs, FPs, FNs for visualization. classifyr2N)rlenncrzerosmatrixrmatches)selfrrrs r.__init__zConfusionMatrix.__init__Es e*6:ii:6Mbhh12SUS[S[]a]d]dgh]hjnjqjqtuju\vSw  )r t r0mtypebatchidxrNc|jy|jD]k\}}|dvr|j||xx||gz cc<)|dk(s/|j||xx|jdkDr |d|dzk(gn||gz cc<my)a Append the matches to TP, FP, FN or GT list for the last batch. This method updates the matches dictionary by appending specific batch data to the appropriate match type (True Positive, False Positive, or False Negative). Args: mtype (str): Match type identifier ('TP', 'FP', 'FN' or 'GT'). batch (Dict[str, Any]): Batch data containing detection results with keys like 'bboxes', 'cls', 'conf', 'keypoints', 'masks'. idx (int): Index of the specific detection to append from the batch. Note: For masks, handles both overlap and non-overlap cases. When masks.max() > 1.0, it indicates overlap_mask=True with shape (1, H, W), otherwise uses direct indexing. N>clsconfbboxes keypointsmasksrrr2)ritemsr9)rrrrkr_s r._append_matcheszConfusionMatrix._append_matchesTs" <<  KKM [DAq:: U#A&!SE(2&g U#A&quuw}1Q437?*;STUXSYRZZ&  [r0predstargetsc2tj|dddftj|}}t|jj |jj D]\}}|j ||xxdz cc<!y)z Update confusion matrix for classification task. Args: preds (List[N, min(nc,5)]): Predicted class labels. targets (List[N, 1]): Ground truth class labels. Nrr2)r7r{zipcpunumpyr)rrrpts r.process_cls_predsz!ConfusionMatrix.process_cls_predsnsx5)!Q$/71Cw ))+W[[]-@-@-BC #DAq KKN1  "  #r0 detectionsr iou_thresc |d|d}}|jOdDcic]}|ttc}|_tt |D]}|j d|||j ddk(} |d| rdnd hvrd n|}t |dd k(} |j d d k(r| s|jDcic]}||||d |kD}}|djj} t| D];\}} |j| |jfxxdz cc<|j d ||=y| rh|jj} t| D];\}}|j|j|fxxdz cc<|j d||=y|jDcic]}||||d |kD}}|jj} |djj} |d}| r t||n t||}tj ||kD}|d j d rtj"tj$|d||d |dfdddffdj'j)}|d j d dkDr||dddfj+ddd}|t-j.|dddfdd}||dddfj+ddd}|t-j.|ddd fdd}nt-j0d}|j d d kD}|j3j5t\}}}t| D]\}}||k(}|rt7|dk(r| ||j9} |j| |fxxdz cc<| |k(r%|j d|||j9t|j d |||j9|j d|||j|j|fxxdz cc<|j d||t| D]J\}} t;||k(r|j| |jfxxdz cc<|j d ||Lycc}wcc}wcc}w)a) Update confusion matrix for object detection task. Args: detections (Dict[str, torch.Tensor]): Dictionary containing detected bounding boxes and their associated information. Should contain 'cls', 'conf', and 'bboxes' keys, where 'bboxes' can be Array[N, 4] for regular boxes or Array[N, 5] for OBB with angle. batch (Dict[str, Any]): Batch dictionary containing ground truth data with 'bboxes' (Array[M, 4]| Array[M, 5]) and 'cls' (Array[M]) keys, where M is the number of ground truth objects. conf (float, optional): Confidence threshold for detections. iou_thres (float, optional): IoU threshold for matching detections to ground truth. rrN>FNFPGTTPrr2g{Gz?MbP?rrrrrr3rHT) return_index)rr)rrlistrangerrshapekeysinttolist enumeraterrrrAr7wherer{stackrrargsortruniquer transposeastypergitemany)rrrrrgt_cls gt_bboxesriis_obbno_preddetection_classesdc gt_classesgcrrrrnm0m1_js r. process_batchzConfusionMatrix.process_batchzs&"%L%/  << #:RSQA{400SDL3v;' 5$$T5!4 5#q(fd%@@tdj'(A- <<?a S]SbSbSdeaaAz&/AD/H!IIe e$.u$5$9$9$;$B$B$D!&'89>EArKKDGG ,1,((z1=>  ,,.J":. 52 DGGRK(A-($$T5!4 5 KU??K\]aaAz&'9D'@AA] ]ZZ\((* &u-113::<H%28mIv.giQW>X KKi ( Q4::a=iiQ!2C!ad OAtG4L MqQUUW]]_Gtzz!}q !'!Q$-"7"7"9$B$"?@!"))GAqDM"Ma"PQ!'!Q$-"7"7"9$B$"?@!"))GAqDM"Ma"PQhhv&G MM! q %%'..s3 Bz* 5EAraASVq[&r!uzz|4 BF#q(#8((z2a5::<H((z2a5::<H((ua8 DGGRK(A-($$T5!4 501 :EArrQw< BK(A-($$T:q9 :mTf^sS*=S/<S4c|jS)zReturn the confusion matrix.)rrs r.rzConfusionMatrix.matrixs {{r0c|jj}|jjd|z }|jdk(r||fS|dd|ddfS)z Return true positives and false positives. Returns: tp (np.ndarray): True positives. fp (np.ndarray): False positives. r2rNrH)rdiagonalrgr)rtpfps r.tp_fpzConfusionMatrix.tp_fpsX[[ ! ! # [[__Q " $99 2BxJCR"Sb'8JJr0imgim_filesave_dirc ^|jsyddlm}ddlm}t t }tgdD]\}}|j|} d| vr4tjdgt| dz|j | d<tjt| d|j |z| d <| jD]} || xx| | z cc<|jD cic])\} } | t| rtj| d n| +}} } |j d k(st|dr||d|d<|d z j#dd|||j%ddddgd|d z t'|j(z |j*ddycc} } w)a Plot grid of GT, TP, FP, FN for each image. Args: img (torch.Tensor): Image to plot onto. im_file (str): Image filename to save visualizations. save_dir (Path): Location to save the visualizations to. Nr2) xyxy2xywh) plot_images)rrrrrrr)rn batch_idxrobbvisualizationsT)parentsexist_okrG)z Ground TruthzFalse PositiveszTrue PositiveszFalse Negativesr)pathsfnamer max_subplots conf_thres)ropsrplottingrrrrr7rprrnonesrrrrmkdirrepeatrnamer) rrrrrrlabelsrrmbatchrr_s r. plot_matcheszConfusionMatrix.plot_matchess|| ")T"!":; 'HAu\\%(FV#!&sec&:J6K.KTWT^T^!_v"'**S1A-B3::"VYZ"ZF; [[] 'q VAY&  '  'EKLLNSDAq!#a&U[[A&a7SSyyE!c&*:&;()9:F8  $ $++D4+H  JJq!Q "Z--W 0B0BB** Ts0.F)zConfusionMatrix plot failure)msg normalizec xddlm}|j|r.|jjdj dddzndz }t j ||dk<|jddd\}}t|jj|j} }|jd k\rPtd |jd z} tdd| } || }|| ddfdd| f}|j| zdz | z} |jd k(r| n| dzx} } d| cxkrd krnn | | k(r|dgznd}t jt!|}tddd| zz }tddd| zz }tddd| zz }tddd| zz }t#j$5t#j&d|j)|ddd}|j*j-d| dkrd|rdnt j.|z}t1|d| D]g\}}t1|d| D]Q\}}|||f}t j2|r#|j5|||r|dn t7|ddd ||kDrd!nd"#Si|j9||d$d%&}dddd'd(|zz}|j;d)|d *|j=d+|d *|j?||d,-|jA||jC||jEd.d/d0d/d01|jEd2d/d0d/d03|dk7r(|jG||d4d5|jI||6d7D]M}|d8k7r|jJ|jMd0jNjJ|jMd0O|jQdd9d:|;tS||jUjWdz }|jY|d?@|j[||r ||yy#1swYxYw)Aav Plot the confusion matrix using matplotlib and save it to a file. Args: normalize (bool, optional): Whether to normalize the confusion matrix. save_dir (str, optional): Directory where the plot will be saved. on_plot (callable, optional): An optional callback to pass plots path and data when they are rendered. rNr2rH& .>g{Gzt?)rx )figsizedr3<rc backgroundauto皙?rxrrignoreBluesnone)cmapvmin interpolationbottom?.2fcenter whiteblack)havafontsizecolorgZd;O?皙?)axfractionpadzConfusion Matrixz NormalizedTrue)r3labelpad Predicted)r3r8rTF)axisr)top labelbottomlabeltopy)r=leftright labelleft labelrightZ)r3rotationr1)r3>r>rBrCr)outlinerHgzG?gGz?)rBrCr>r) rz.pngdpi).matplotlib.pyplotpyplotrrgreshapernansubplotsrrvaluesrr9slicerarangerwarningscatch_warnings simplefilterimshowxaxisset_label_positionnanmaxrisnantextrcolorbar set_xlabel set_ylabel set_title set_xticks set_yticks tick_paramsset_xticklabelsset_yticklabelsspines set_visibler6subplots_adjustrlowerreplacesavefigclose)rrron_plotpltarrayfigr6rrrkeep_idxrnn ticklabelsxy_ticks tick_fontsizelabel_fontsizetitle_fontsizebtmimcolor_thresholdrrowrvalcbartitles plot_fnames r.plotzConfusionMatrix.plots ( Y  2 : :1b AD H\]^!vveem,,q!W,5R ))+,dggq 77c>Atww"}%AT4+H(OE(A+&q({3E1qQ&AyyJ.!AE9R23b+2+B"He|n,SY 99S_-ArC"H}- QS2X .QS2X .#tebj()  $ $ & E  ! !( +5wSOB HH ' ' 1Bw"&y!bii>N"O'cr 3 FAs"+CH"5 3#AqDk88C=$,5s3ic#hZ''%'-0?-B'   <<rEt<DD) E*#]Y%>> f~ C kNR H U^ < h h C%TTYZ Cd%4TYZ     zMBS[  \   zM  B> 1AI~ ! ((/ GGNN1  ) )% 0 1 $DE(^)>)>sC)H(I&NN  JC ( #  J  Q E Es .C4P//P9c t|jjdD]A}tjdj t t|j|Cy)z*Print the confusion matrix to the console.rrIN)rrrr infojoinmapstrrrs r.printzConfusionMatrix.printKsIt{{((+, c} } Scc} wcc} } w) a Generate a summarized representation of the confusion matrix as a list of dictionaries, with optional normalization. This is useful for exporting the matrix to various formats such as CSV, XML, HTML, JSON, or SQL. Args: normalize (bool): Whether to normalize the confusion matrix values. decimals (int): Number of decimal places to round the output values to. Returns: (List[Dict[str, float]]): A list of dictionaries, each representing one predicted class with corresponding values for all actual classes. Examples: >>> results = model.val(data="coco8.yaml", plots=True) >>> cm_dict = results.confusion_matrix.summary(normalize=True, decimals=5) >>> print(cm_dict) rNrrz [^a-zA-Z0-9_]rr2rHrr;)rerrrrRsetsubrjaddappendrrgrOroundrrdict)rrrrr clean_namesseenr clean_nameoriginal_cleancounterrprrs r.summaryzConfusionMatrix.summaryPs" -1YY*-DTZZ&&()$tzzO`O`ObJcgsftJtT  +D 0#t5A;3N2p q  2p s(F6FFF)rr+)TN)Fr)__name__ __module__ __qualname____doc__rrrboolrrrrr7Tensorrr4rrrrrrrrr rrrrrr0r.rr9s 02x^c 4d38n 4 4W[ 4[S[c3h[c[d[4 #tELL'9 #DDV #[_ #  N:ell*+N:CH~N: N:  N:  N:` KuRZZ34 K$  $ s$ d$ t$ L12^F dF SF 3F P< " " " T$sTYzJZE[" r0rrAfctt||zdzdzdz}tj|dz}tj||dz|||dzfd}tj |tj||z dS)zBox filter of fraction f.r3r2rrHvalid)mode)rrrr  concatenateconvolve)rArnfryps r.smoothrus{ s1vzA~ ! #a 'B aA QqT1a!B%i0! 4B ;;r2772;+' ::r0z pr_curve.pngpxpyaprrc ddlm}|jdddd\}}tj|d}dt |cxkrdkrDnnAt |jD](\} } |j|| d|| d || dfd  *n|j||dd |j||jdddd|dddfjd d|jd|jd|jdd|jdd|jdd|jd|j!|d|j#||r ||yy)a Plot precision-recall curve. Args: px (np.ndarray): X values for the PR curve. py (np.ndarray): Y values for the PR curve. ap (np.ndarray): Average precision values. save_dir (Path, optional): Path to save the plot. names (Dict[int, str], optional): Dictionary mapping class indices to class names. on_plot (callable, optional): Function to call after plot is saved. rNr2rrTr tight_layout)r=rI.3f linewidthlabelgreyrr4rblue all classes z mAP@0.5rr4rRecall Precisiongp= ף?r2 upper leftbbox_to_anchorloczPrecision-Recall CurverJrK)rMrNrQrrrrrrmeanr_r`set_xlimset_ylimlegendrarlrm) rrrrrrnrorqr6rrAs r. plot_pr_curver}sY($ll1adlCGC "1 B3u:bddO LDAq GGBQq !Bq!tHS>.JG K L B!62GGB av|BqRStHMMO\_K``h=iGjMM(MM+KK1KK1IIYLI9LL)*KKcK"IIcNr0z mc_curve.png ConfidenceMetricxlabelylabelc ddlm}|jdddd\}} dt|cxkrdkr0nn-t |D]\} } | j || d||  n| j ||j dd t|jdd } | j || d d d| jdd|| jd| j|| j|| jdd| jdd| jdd| j!|d|j#|d|j%||r ||yy)a Plot metric-confidence curve. Args: px (np.ndarray): X values for the metric-confidence curve. py (np.ndarray): Y values for the metric-confidence curve. save_dir (Path, optional): Path to save the plot. names (Dict[int, str], optional): Dictionary mapping class indices to class names. xlabel (str, optional): X-axis label. ylabel (str, optional): Y-axis label. on_plot (callable, optional): Function to call after plot is saved. rNr2rTrrrrrr!rrrr,z at rrrrrz-Confidence CurverJrK)rMrNrQrrrrrrr9argmaxr_r`rrrrarlrm) rrrrrrrnrorqr6rrAs r. plot_mc_curversS,$ll1adlCGC3u:bM =DAq GGBQq G < = BDDAV4rwwqz3AGGBQfl15573-tTVWXW_W_WaTbcfSg4hGiMM&MM&KK1KK1IIYLI9LLF8,-.KKcK"IIcNr0recall precisionctjdg|dgf}tjdg|dgf}tjtjj tj|}d}|dk(rztj ddd}t jtjdrtjntj}|tj||||}nKtj|dd|dd k7d}tj||dz||z ||dzz}|||fS) a Compute the average precision (AP) given the recall and precision curves. Args: recall (list): The recall curve. precision (list): The precision curve. Returns: ap (float): Average precision. mpre (np.ndarray): Precision envelope curve. mrec (np.ndarray): Modified recall curve with sentinel values added at the beginning and end. r$rinterprr2ez>=2.0NrH)rrflipr! accumulatelinspacer check_version __version__ trapezoidtrapzrrrg) rrmrecmpremethodrfuncrrs r. compute_aprs >>C5&3%0 1D >>C5)cU3 4D 772::((7 8DF  KK1c "%33BNNGLr||RTRZRZ "))AtT*A . HHT!"Xcr* +A . VVT!a%[47*d1q5k9 : tT>r0gؗҜ*=A Mfeeg !8sax 2a5y  #ell1oc "iiT!WHfQTlC 39% iiT!WHi1oAF rxx{# =A$.vad|Yq!t_$M !Br1uItTAv""299Qd#;< =)=2,7"((;'BHHY MTQ!u*Qa[ ME M ab(x|5L*Le]dea8 .E#EuUYcjka(x{-C"CUS^hopa(x{-C"CUS[elmx}}Q%,,.Aq!t}gadmXad^"qA b& B C.2  $ $ &B r1aR!6!6s!;WgxYZ\g gg Ns 6 L=L=ceZdZdZddZedeeje ffdZ edeeje ffdZ ede fdZ ede fdZede fd Zede fd Zede fd Zde e fd Zd edee e e e ffdZedejfdZde fdZdefdZede fdZede e fdZy)ra Class for computing evaluation metrics for Ultralytics YOLO models. Attributes: p (list): Precision for each class. Shape: (nc,). r (list): Recall for each class. Shape: (nc,). f1 (list): F1 score for each class. Shape: (nc,). all_ap (list): AP scores for all classes and all IoU thresholds. Shape: (nc, 10). ap_class_index (list): Index of class for each AP score. Shape: (nc,). nc (int): Number of classes. Methods: ap50: AP at IoU threshold of 0.5 for all classes. ap: AP at IoU thresholds from 0.5 to 0.95 for all classes. mp: Mean precision of all classes. mr: Mean recall of all classes. map50: Mean AP at IoU threshold of 0.5 for all classes. map75: Mean AP at IoU threshold of 0.75 for all classes. map: Mean AP at IoU thresholds from 0.5 to 0.95 for all classes. mean_results: Mean of results, returns mp, mr, map50, map. class_result: Class-aware result, returns p[i], r[i], ap50[i], ap[i]. maps: mAP of each class. fitness: Model fitness as a weighted combination of metrics. update: Update metric attributes with new evaluation results. curves: Provides a list of curves for accessing specific metrics like precision, recall, F1, etc. curves_results: Provide a list of results for accessing specific metrics like precision, recall, F1, etc. rNcXg|_g|_g|_g|_g|_d|_y)zSInitialize a Metric instance for computing evaluation metrics for the YOLOv8 model.rN)rrrall_apap_class_indexrrs r.rzMetric.__init__vs-  r0cVt|jr|jdddfSgS)z Return the Average Precision (AP) at an IoU threshold of 0.5 for all classes. Returns: (np.ndarray | list): Array of shape (nc,) with AP50 values per class, or an empty list if not available. Nr)rrrs r.ap50z Metric.ap50s'%( $4t{{1a4 <"B>r0cdt|jr|jjSdS)z Return the Mean Precision of all classes. Returns: (float): The mean precision of all classes. r$)rrrrs r.mpz Metric.mp#!$DFF tvv{{}44r0cdt|jr|jjSdS)z Return the Mean Recall of all classes. Returns: (float): The mean recall of all classes. r$)rrrrs r.mrz Metric.mrrr0crt|jr!|jdddfjSdS)z Return the mean Average Precision (mAP) at an IoU threshold of 0.5. Returns: (float): The mAP at an IoU threshold of 0.5. Nrr$rrs r.map50z Metric.map500,/t{{+;t{{1a4 %%'DDr0crt|jr!|jdddfjSdS)z Return the mean Average Precision (mAP) at an IoU threshold of 0.75. Returns: (float): The mAP at an IoU threshold of 0.75. Nrr$rrs r.map75z Metric.map75rr0cdt|jr|jjSdS)z Return the mean Average Precision (mAP) over IoU thresholds of 0.5 - 0.95 in steps of 0.05. Returns: (float): The mAP over IoU thresholds of 0.5 - 0.95 in steps of 0.05. r$rrs r.rz Metric.maps'&)%5t{{!>3>r0c^|j|j|j|jgS)z+Return mean of results, mp, mr, map50, map.)rrrrrs r. mean_resultszMetric.mean_resultss!$**dhh77r0rcv|j||j||j||j|fS)z6Return class-aware result, p[i], r[i], ap50[i], ap[i].)rrrrrs r. class_resultzMetric.class_results1vvay$&&)TYYq\4771:==r0ctj|j|jz}t |j D]\}}|j |||<|S)zReturn mAP of each class.)rrrrrrr)rmapsrrs r.r z Metric.mapssSxx 488+d112 !DAqggajDG ! r0cgd}tjtj|j|zj S)z:Return model fitness as a weighted combination of metrics.)r$r$r!g?)r nan_to_numrpr rg)rws r.fitnesszMetric.fitnesss6  bhht'8'8':;Return a list of curves for accessing specific metrics curves.rrs r.curvesz Metric.curves  r0c|j|jddg|j|jddg|j|jddg|j|jddggS)rrrrr)rrrrrrs r.curves_resultszMetric.curves_resultss`WWd&&+ > WWdmm\4 8 WWdllL+ > WWdllL( ;   r0rN)rrrrrpropertyr rrrrrr4rrrrrr rrr r rtuplerrrrr0r.rrYs8=eBJJ,-==?E"**d*+??5E555E55EuEEEuEE?U??8d5k8>c>eE5%,F&G>bjjHH e< T   r0c eZdZdZifdeeefddfdZdeeefddfdZ e dd dfd e d e deee jffd Zd ZedeefdZdeefdZdedeeeeeffdZede jfdZedefdZedefdZedeeeffdZedeefdZedeefdZdde dedeeeeffdZy) DetMetricsa Utility class for computing detection metrics such as precision, recall, and mean average precision (mAP). Attributes: names (Dict[int, str]): A dictionary of class names. box (Metric): An instance of the Metric class for storing detection results. speed (Dict[str, float]): A dictionary for storing execution times of different parts of the detection process. task (str): The task type, set to 'detect'. stats (Dict[str, List]): A dictionary containing lists for true positives, confidence scores, predicted classes, target classes, and target images. nt_per_class: Number of targets per class. nt_per_image: Number of targets per image. Methods: update_stats: Update statistics by appending new values to existing stat collections. process: Process predicted results for object detection and update metrics. clear_stats: Clear the stored statistics. keys: Return a list of keys for accessing specific metrics. mean_results: Calculate mean of detected objects & return precision, recall, mAP50, and mAP50-95. class_result: Return the result of evaluating the performance of an object detection model on a specific class. maps: Return mean Average Precision (mAP) scores per class. fitness: Return the fitness of box object. ap_class_index: Return the average precision index per class. results_dict: Return dictionary of computed performance metrics and statistics. curves: Return a list of curves for accessing specific metrics curves. curves_results: Return a list of computed performance metrics and statistics. summary: Generate a summarized representation of per-class detection metrics as a list of dictionaries. rrNc||_t|_ddddd|_d|_t ggggg|_d|_d|_y)z Initialize a DetMetrics instance with a save directory, plot flag, and class names. Args: names (Dict[int, str], optional): Dictionary of class names. r$ preprocess inferenceloss postprocessr)rrrr target_imgN) rrboxspeedrrstats nt_per_class nt_per_imagerrs r.rzDetMetrics.__init__%sQ 8$'c3WZ[  Rb2"QST   r0statc|jjD]#}|j|j||%y)a  Update statistics by appending new values to existing stat collections. Args: stat (Dict[str, any]): Dictionary containing new statistical values to append. Keys should match existing keys in self.stats. N)r*rr)rr.rs r. update_statszDetMetrics.update_stats4s8" *A JJqM a ) *r0.Frrc |jjDcic]\}}|tj|d}}}t |dk(r|St |d|d|d|d|||j |d dd }t |j |j_|jj|tj|djtt |j  |_ tj|d jtt |j  |_|Scc}}w) a Process predicted results for object detection and update metrics. Args: save_dir (Path): Directory to save plots. Defaults to Path("."). plot (bool): Whether to plot precision-recall curves. Defaults to False. on_plot (callable, optional): Function to call after plots are generated. Defaults to None. Returns: (Dict[str, np.ndarray]): Dictionary containing concatenated statistics arrays. rrrrrBox)rrrrnrr3N) minlengthr')r*rrrrrrr(rrbincountrrr+r,)rrrrnrr_r*rs r.processzDetMetrics.process?s6:ZZ5E5E5GHTQBNN1a((HH u:?L $K &M *  , **   " $**o   KKl(;(B(B3(GSVW[WaWaSbcKKl(;(B(B3(GSVW[WaWaSbc %Is!E cb|jjD]}|jy)zClear the stored statistics.N)r*rRclear)rr_s r. clear_statszDetMetrics.clear_stats_s'""$ A GGI r0c gdS)z5Return a list of keys for accessing specific metrics.)zmetrics/precision(B)zmetrics/recall(B)zmetrics/mAP50(B)zmetrics/mAP50-95(B)rrs r.rzDetMetrics.keysds hgr0c6|jjS)zSCalculate mean of detected objects & return precision, recall, mAP50, and mAP50-95.)r(r rs r.r zDetMetrics.mean_resultsisxx$$&&r0rc8|jj|S)zaReturn the result of evaluating the performance of an object detection model on a specific class.)r(r rs r.r zDetMetrics.class_resultmsxx$$Q''r0c.|jjS)z5Return mean Average Precision (mAP) scores per class.)r(r rs r.r zDetMetrics.mapsqsxx}}r0c6|jjS)z!Return the fitness of box object.)r(rrs r.rzDetMetrics.fitnessvsxx!!r0c.|jjS)z-Return the average precision index per class.)r(rrs r.rzDetMetrics.ap_class_index{xx&&&r0ctt|jdgz|j|jgzS)zAReturn dictionary of computed performance metrics and statistics.r)rrrr rrs r. results_dictzDetMetrics.results_dicts5C YK/1B1B1D ~1UVWWr0c gdS)r)Precision-Recall(B)F1-Confidence(B)Precision-Confidence(B)Recall-Confidence(B)rrs r.rzDetMetrics.curvess nmr0c.|jjSz=Return a list of computed performance metrics and statistics.)r(rrs r.rzDetMetrics.curves_resultsr@r0rrcr|jj|jj|jjd}t t |dDcgc]}|j |j||j|j||j|j|d|jDcic]\}}|t|||c}}t|j|d|t|j|d|dc}}}Scc}}wcc}}}w)a Generate a summarized representation of per-class detection metrics as a list of dictionaries. Includes shared scalar metrics (mAP, mAP50, mAP75) alongside precision, recall, and F1-score for each class. Args: normalize (bool): For Detect metrics, everything is normalized by default [0-1]. decimals (int): Number of decimal places to round the metrics values to. Returns: (List[Dict[str, Any]]): A list of dictionaries, each representing one class with corresponding metric values. Examples: >>> results = model.val(data="coco8.yaml") >>> detection_summary = results.summary() >>> print(detection_summary) )Box-PzBox-RzBox-F1rK)ClassImages Instancesr3r)mAP50zmAP50-95) r(rrrrrrrr,r+rrr )rrr per_classrrr_s r.rzDetMetrics.summarys!$XXZZXXZZhhkk 3y123   D$7$7$:;++D,?,?,BC!..t/B/B1/EF 9B8IJ11eAaD(++J   t003A6A!$"3"3A"6q"98D    K  sA+D2 D,#AD2,D2Tr)rrrrrrrrrr0rrrrr6r9rrrr4r rr r rrrBrrrrr0r.r r s802 !d38n !d ! *c3h *D *(,CyudW[\_acakak\kWl@ hd3ihh'd5k'(c(eE5%,F&G(bjj"""'''Xd3:.XXnS nn'T ''      DcSVhDX  r0r c VeZdZdZifdeeefddfdZedddfded e deee jffd Z e deefd Zdeefd Zd edeefdZe de jfdZe defdZe deefdZe deefdZdde dedeeeeffdZy)SegmentMetricsa. Calculate and aggregate detection and segmentation metrics over a given set of classes. Attributes: names (Dict[int, str]): Dictionary of class names. box (Metric): An instance of the Metric class for storing detection results. seg (Metric): An instance of the Metric class to calculate mask segmentation metrics. speed (Dict[str, float]): A dictionary for storing execution times of different parts of the detection process. task (str): The task type, set to 'segment'. stats (Dict[str, List]): A dictionary containing lists for true positives, confidence scores, predicted classes, target classes, and target images. nt_per_class: Number of targets per class. nt_per_image: Number of targets per image. Methods: process: Process the detection and segmentation metrics over the given set of predictions. keys: Return a list of keys for accessing metrics. mean_results: Return the mean metrics for bounding box and segmentation results. class_result: Return classification results for a specified class index. maps: Return mAP scores for object detection and semantic segmentation models. fitness: Return the fitness score for both segmentation and bounding box models. curves: Return a list of curves for accessing specific metrics curves. curves_results: Provide a list of computed performance metrics and statistics. summary: Generate a summarized representation of per-class segmentation metrics as a list of dictionaries. rrNcztj||t|_d|_g|j d<y)z Initialize a SegmentMetrics instance with a save directory, plot flag, and class names. Args: names (Dict[int, str], optional): Dictionary of class names. segmenttp_mN)r rrsegrr*r-s r.rzSegmentMetrics.__init__s2 D%(8  6r0r1Frrc tj||||}t|d|d|d|d||||jd dd }t |j|j _|j j||S) a Process the detection and segmentation metrics over the given set of predictions. Args: save_dir (Path): Directory to save plots. Defaults to Path("."). plot (bool): Whether to plot precision-recall curves. Defaults to False. on_plot (callable, optional): Function to call after plots are generated. Defaults to None. Returns: (Dict[str, np.ndarray]): Dictionary containing concatenated statistics arrays. rrVrrrMaskrrnrrrr3N)r r6rrrrWrr)rrrrnr* results_masks r.r6zSegmentMetrics.processs""44"I# &M &M *  , **   "  $**o   % r0cJtjj|gdzS)z,Return a list of keys for accessing metrics.)zmetrics/precision(M)zmetrics/recall(M)zmetrics/mAP50(M)zmetrics/mAP50-95(M)r rfgetrs r.rzSegmentMetrics.keys%##D)-   r0cbtj||jjzS)zBReturn the mean metrics for bounding box and segmentation results.)r r rWrs r.r zSegmentMetrics.mean_resultss%&&t,txx/D/D/FFFr0rcftj|||jj|zS)z:Return classification results for a specified class index.)r r rWrs r.r zSegmentMetrics.class_results)&&tQ/$((2G2G2JJJr0cntjj||jjzS)zHReturn mAP scores for object detection and semantic segmentation models.)r r r^rWrs r.r zSegmentMetrics.mapss&##D)DHHMM99r0cv|jjtjj|zS)zGReturn the fitness score for both segmentation and bounding box models.)rWrr r^rs r.rzSegmentMetrics.fitness s-xx!J$6$6$;$;D$AAAr0cJtjj|gdzS)r)zPrecision-Recall(M)zF1-Confidence(M)zPrecision-Confidence(M)zRecall-Confidence(M)r rr^rs r.rzSegmentMetrics.curvess'  %%d+/   r0cntjj||jjzSrI)r rr^rWrs r.rzSegmentMetrics.curves_resultss*((--d3dhh6M6MMMr0rrc n|jj|jj|jjd}tj |||}t |D]G\}}|ji|jDcic]\}}|t|||c}}I|Scc}}w)a Generate a summarized representation of per-class segmentation metrics as a list of dictionaries. Includes both box and mask scalar metrics (mAP, mAP50, mAP75) alongside precision, recall, and F1-score for each class. Args: normalize (bool): For Segment metrics, everything is normalized by default [0-1]. decimals (int): Number of decimal places to round the metrics values to. Returns: (List[Dict[str, Any]]): A list of dictionaries, each representing one class with corresponding metric values. Examples: >>> results = model.val(data="coco8-seg.yaml") >>> seg_summary = results.summary(decimals=4) >>> print(seg_summary) )zMask-PzMask-RzMask-F1) rWrrrr rrrrr rrrrPrrrrr_s r.rzSegmentMetrics.summarys$hhjjhhjjxx{{ $$T9h?g& SDAq HHQioo>OPdaE!A$11PQ R SQ B1rQ)rrrrrrrrrrrrr6rrrr4r r r rrrrrrr0r.rSrSsR202  d38n  d  (,CyudW[\_acakak\kWl8 d3i  Gd5kGKcKd5kK:bjj::BBB S   NT NNDcSVhDXr0rSc beZdZdZifdeeefddffd Zedddfded e deee jffd Z e deefd Zdeefd Zd edeefdZe de jfdZe defdZe deefdZe deefdZdde dedeeeeffdZxZS) PoseMetricsa Calculate and aggregate detection and pose metrics over a given set of classes. Attributes: names (Dict[int, str]): Dictionary of class names. pose (Metric): An instance of the Metric class to calculate pose metrics. box (Metric): An instance of the Metric class for storing detection results. speed (Dict[str, float]): A dictionary for storing execution times of different parts of the detection process. task (str): The task type, set to 'pose'. stats (Dict[str, List]): A dictionary containing lists for true positives, confidence scores, predicted classes, target classes, and target images. nt_per_class: Number of targets per class. nt_per_image: Number of targets per image. Methods: process: Process the detection and pose metrics over the given set of predictions. R keys: Return a list of keys for accessing metrics. mean_results: Return the mean results of box and pose. class_result: Return the class-wise detection results for a specific class i. maps: Return the mean average precision (mAP) per class for both box and pose detections. fitness: Return combined fitness score for pose and box detection. curves: Return a list of curves for accessing specific metrics curves. curves_results: Provide a list of computed performance metrics and statistics. summary: Generate a summarized representation of per-class pose metrics as a list of dictionaries. rrNcnt||t|_d|_g|j d<y)z Initialize the PoseMetrics class with directory path, class names, and plotting options. Args: names (Dict[int, str], optional): Dictionary of class names. posetp_pN)superrrrmrr*)rr __class__s r.rzPoseMetrics.__init__Us0 H   6r0r1Frrc tj||||}t|d|d|d|d||||jd dd }t |j|j _|j j||S) a Process the detection and pose metrics over the given set of predictions. Args: save_dir (Path): Directory to save plots. Defaults to Path("."). plot (bool): Whether to plot precision-recall curves. Defaults to False. on_plot (callable, optional): Function to call after plots are generated. Returns: (Dict[str, np.ndarray]): Dictionary containing concatenated statistics arrays. rrnrrrPoserZr3N)r r6rrrrmrr)rrrrnr* results_poses r.r6zPoseMetrics.processas""44"I# &M &M *  , **   "  4::  & r0cJtjj|gdzS)z(Return a list of evaluation metric keys.)zmetrics/precision(P)zmetrics/recall(P)zmetrics/mAP50(P)zmetrics/mAP50-95(P)r]rs r.rzPoseMetrics.keys}r_r0cbtj||jjzS)z(Return the mean results of box and pose.)r r rmrs r.r zPoseMetrics.mean_resultss%&&t,tyy/E/E/GGGr0rcftj|||jj|zS)z?Return the class-wise detection results for a specific class i.)r r rmrs r.r zPoseMetrics.class_results)&&tQ/$))2H2H2KKKr0cntjj||jjzS)zSReturn the mean average precision (mAP) per class for both box and pose detections.)r r r^rmrs r.r zPoseMetrics.mapss&##D)DIINN::r0cv|jjtjj|zS)z9Return combined fitness score for pose and box detection.)rmrr r^rs r.rzPoseMetrics.fitnesss-yy  "Z%7%7%<%>> results = model.val(data="coco8-pose.yaml") >>> pose_summary = results.summary(decimals=4) >>> print(pose_summary) )zPose-PzPose-RzPose-F1) rmrrrr rrrrrrhs r.rzPoseMetrics.summarys$iikkiikkyy|| $$T9h?g& SDAq HHQioo>OPdaE!A$11PQ R SQrirQ)rrrrrrrrrrrrr6rrrr4r r r rrrrr __classcell__)rps@r.rkrk;sS202  d38n  d  (,CyudW[\_acakak\kWl8 d3i  Hd5kHLcLd5kL;bjj;;CCC  S    OT OODcSVhDXr0rkc eZdZdZddZdej dej fdZede fdZ ede e e ffd Z edee fd Zedefd Zedefd Zdd ededee e e ffdZy)ClassifyMetricsa Class for computing classification metrics including top-1 and top-5 accuracy. Attributes: top1 (float): The top-1 accuracy. top5 (float): The top-5 accuracy. speed (dict): A dictionary containing the time taken for each step in the pipeline. task (str): The task type, set to 'classify'. Methods: process: Process target classes and predicted classes to compute metrics. fitness: Return mean of top-1 and top-5 accuracies as fitness score. results_dict: Return a dictionary with model's performance metrics and fitness score. keys: Return a list of keys for the results_dict property. curves: Return a list of curves for accessing specific metrics curves. curves_results: Provide a list of computed performance metrics and statistics. summary: Generate a single-row summary of classification metrics (Top-1 and Top-5 accuracy). rNcFd|_d|_ddddd|_d|_y)z&Initialize a ClassifyMetrics instance.rr$r"rN)top1top5r)rrs r.rzClassifyMetrics.__init__s(  $'c3WZ[  r0rpredcVtj|tj|}}|dddf|k(j}tj|dddf|j dj fd}|j dj\|_|_ y)z Process target classes and predicted classes to compute metrics. Args: targets (torch.Tensor): Target classes. pred (torch.Tensor): Predicted classes. Nrr2ry) r7r{r4rr9rRrrrr)rrrcorrectaccs r.r6zClassifyMetrics.processs $7);g1d7#t+224kk71a4='++a.*?*?@aH"xx{113 49r0c:|j|jzdz S)z;Return mean of top-1 and top-5 accuracies as fitness score.r3)rrrs r.rzClassifyMetrics.fitnesss DII%**r0ctt|jdgz|j|j|j gS)zGReturn a dictionary with model's performance metrics and fitness score.r)rrrrrrrs r.rBzClassifyMetrics.results_dicts4C YK/$))TYY 1UVWWr0c ddgS)z4Return a list of keys for the results_dict property.zmetrics/accuracy_top1zmetrics/accuracy_top5rrs r.rzClassifyMetrics.keyss()@AAr0cgSrrrs r.rzClassifyMetrics.curvesrr0cgSrrrs r.rzClassifyMetrics.curves_resultsrr0rrc^t|j|t|j|dgS)aa Generate a single-row summary of classification metrics (Top-1 and Top-5 accuracy). Args: normalize (bool): For Classify metrics, everything is normalized by default [0-1]. decimals (int): Number of decimal places to round the metrics values to. Returns: (List[Dict[str, float]]): A list with one dictionary containing Top-1 and Top-5 classification accuracy. Examples: >>> results = model.val(data="imagenet10") >>> classify_summary = results.summary(decimals=4) >>> print(classify_summary) )top1_acctop5_acc)rrr)rrrs r.rzClassifyMetrics.summary s) #499h7U499V^E_`aar0rrQ)rrrrrr7rr6rr4rrrrBrrrrrrrrr0r.r~r~s& 4u|| 45<< 4+++Xd3:.XXBd3iBBbbbDcSXjIYDZbr0r~c.eZdZdZifdeeefddfdZy) OBBMetricsa Metrics for evaluating oriented bounding box (OBB) detection. Attributes: names (Dict[int, str]): Dictionary of class names. box (Metric): An instance of the Metric class for storing detection results. speed (Dict[str, float]): A dictionary for storing execution times of different parts of the detection process. task (str): The task type, set to 'obb'. stats (Dict[str, List]): A dictionary containing lists for true positives, confidence scores, predicted classes, target classes, and target images. nt_per_class: Number of targets per class. nt_per_image: Number of targets per image. References: https://arxiv.org/pdf/2106.06072.pdf rrNc>tj||d|_y)z Initialize an OBBMetrics instance with directory, plotting, and class names. Args: names (Dict[int, str], optional): Dictionary of class names. rN)r rrr-s r.rzOBBMetrics.__init__-s D%( r0)rrrrrrrrrr0r.rrs' 02 d38n d r0r)FHz>)r)TFFFr)r!)r5)4rrJrU collectionsrpathlibrtypingrrrrr rrr7ultralytics.utilsr r r r rrrp OKS_SIGMArrr4r/rrArbrirurrrrrrrrrrrrrr rSrkr~rrr0r.rs #00 cc BHH st   %2::%RZZ%d%%Z\ZdZd%@I%,,IellII%,,I6D ,,D ,,D D  D  D  D D \\DN(ELL((E(U\\((bfX ,,X#llX27,,XGKE{XY^X \\X0I%,,I5u||UZUaUa9a3bI(+%,,+ell+$+U+^c^j^j+^`d$  bjj( )$16u||RZZ7O1P$W\$ \\$N&E&E%,$7&"y oy x ;bjj;U;bjj; .)  ( ( ( ( ( S> ((V.) * * ** S> *  *  **ZtE{tE{uUBJJXZXbXbEb?cL V^h ^h **^hjj^h ^h  ^h^h S>^h ^h ^h ^hBl [l ^g og TFZFRJ*JZQbk?Qbhr0