|hXddlZddlZddlZddlZddlmZddlZddlZddl Z ddl m cm Z ddlmZddlmZGddej$ZdEdedefd ZdFd ed efd Zd efdZdGdedefdZ dHdededededededededededededefdZd Zd!ZdId"Zd#Zd$Z dJd%ed&ed'ed(efd)Z!dKd%ed&ed*ed+efd,Z"d-Z#d.Z$d/Z%d0Z&d1Z'd2Z(d3Z)dLd4efd5Z*d6Z+dMd7efd8Z,d9Z-dNd efd:Z.dOd;ed efd<Z/d=Z0dPd>e1fd?Z2d@e jfdAejhfdBZ5dCZ6dDZ7y)QN)Optional)LOGGER) batch_probioucTeZdZdZd dedeejfdZdZ dZ dZ d Z y) Profilea< Ultralytics Profile class for timing code execution. Use as a decorator with @Profile() or as a context manager with 'with Profile():'. Provides accurate timing measurements with CUDA synchronization support for GPU operations. Attributes: t (float): Accumulated time in seconds. device (torch.device): Device used for model inference. cuda (bool): Whether CUDA is being used for timing synchronization. Examples: Use as a context manager to time code execution >>> with Profile(device=device) as dt: ... pass # slow operation here >>> print(dt) # prints "Elapsed time is 9.5367431640625e-07 s" Use as a decorator to time function execution >>> @Profile() ... def slow_function(): ... time.sleep(0.1) Ntdevicecx||_||_t|xrt|j d|_y)z Initialize the Profile class. Args: t (float): Initial accumulated time in seconds. device (torch.device, optional): Device used for model inference to enable CUDA synchronization. cudaN)rr boolstr startswithr )selfrr s T/var/www/html/test/engine/venv/lib/python3.12/site-packages/ultralytics/utils/ops.py__init__zProfile.__init__*s2 BCK$:$:6$BC c0|j|_|S)z Start timing.)timestartrs r __enter__zProfile.__enter__6sYY[  rc|j|jz |_|xj|jz c_y)z Stop timing.N)rrdtr)rtypevalue tracebacks r__exit__zProfile.__exit__;s*))+ * $''rc"d|jdS)zIReturn a human-readable string representing the accumulated elapsed time.zElapsed time is z s)rrs r__str__zProfile.__str__@s!$&&,,rc|jr)tjj|jt j S)z9Get current time with CUDA synchronization if applicable.)r torch synchronizer r perf_counterrs rrz Profile.timeDs. 99 JJ " "4;; /  ""r)N) __name__ __module__ __qualname____doc__floatrr!r rrrrrrrrrs:. D% Dx /E D  -#rrwidthheightc|j\}}tj|jdk|jdk|j |kD|j |kDgj dk\r$|j d|}|j d|}|dk\|dk\z||kz||kz}||}||}t|r]tj|j|j|j |j g|jStjd|jS)a+ Convert segment coordinates to bounding box coordinates. Converts a single segment label to a box label by finding the minimum and maximum x and y coordinates. Applies inside-image constraint and clips coordinates when necessary. Args: segment (torch.Tensor): Segment coordinates in format (N, 2) where N is number of points. width (int): Width of the image in pixels. height (int): Height of the image in pixels. Returns: (np.ndarray): Bounding box coordinates in xyxy format [x1, y1, x2, y2]. rdtype) Tnparrayminmaxsumclipanyr0zeros)segmentr+r,xyinsides r segment2boxr?Ks 99DAq xx1aeegk1557U?AEEGftjtjtj tj fn=tjtjtjtj f\}}}}dddf}fdtddD\}}}||||} } |dz | z|dz | zg} | dz | z|dz | zg} || d} || d} || z| z} || z| z }|| z | z }|| z | z}|| |||gdS)a Convert batched Oriented Bounding Boxes (OBB) from [xywh, rotation] to [xy1, xy2, xy3, xy4] format. Args: x (numpy.ndarray | torch.Tensor): Boxes in [cx, cy, w, h, rotation] format with shape (N, 5) or (B, N, 5). Rotation values should be in radians from 0 to pi/2. Returns: (numpy.ndarray | torch.Tensor): Converted corner points with shape (N, 4, 2) or (B, N, 4, 2). .NrDc36K|]}d||dzfyw).rCNr*).0rr<s r z!xywhr2xyxyxyxy..[s":1S!a!e)^$:srrar) rTr!rUcossinrrr3 concatenater)r<rrrrctrrrr cos_value sin_valuevec1vec2pt1pt2pt3pt4s` rxywhr2xyxyxyxyrIs2 a & EIIuyy%++6ffbffbnnbhh 7Cc5 C!G*C:eAqk:KAq%u:s5zyI EI q1uy0 1D BFY A 1 2D tR=D tR=D *t C *t C *t C *t C #sC%r **rct|tjr|jnt j |}|d|dz|d<|d|dz|d<|S)a Convert bounding box from [x1, y1, w, h] to [x1, y1, x2, y2] where xy1=top-left, xy2=bottom-right. Args: x (np.ndarray | torch.Tensor): Input bounding box coordinates. Returns: (np.ndarray | torch.Tensor): Bounding box coordinates in xyxy format. rHrFrIrGrrs r ltwh2xyxyr hrrcg}|D]^}|j\}}|j|j|j|j|jg`t t j |S)a% Convert segment labels to box labels, i.e. (cls, xy1, xy2, ...) to (cls, xywh). Args: segments (list): List of segments where each segment is a list of points, each point is [x, y] coordinates. Returns: (np.ndarray): Bounding box coordinates in xywh format. )r2rr5r6rr3r4)segmentsrMsr<r=s rsegments2boxesr xsg E ;ss1 aeegquuw9:; RXXe_ %%rrqc t|D]D\}}t||k(rtj||ddddffd}tjdt|dz t||kr|t|z n|}tj t|}t||kr+tj |tj|||n|}tjtdDcgc] }tj|||dd|f"c}tjjddj|<G|Scc}w)a7 Resample segments to n points each using linear interpolation. Args: segments (list): List of (N, 2) arrays where N is the number of points in each segment. n (int): Number of points to resample each segment to. Returns: (list): Resampled segments with n points each. rrCN)axisrDr/ra) rrr3rlinspacerhinsert searchsortedrinterpfloat32rr2)r rqrr r<xps rresample_segmentsrs(#  1 q6Q;  NNAq1ay> 2 KK3q6A:SVaZq3q6zQ G YYs1v 8;A BIIaB/ 4 NNuQxH!BIIaQq!tW5HPRPZPZ [ c cdegi j l l     OIs8%E cr|j\}}}tj|dddddfdd\}}}}tj||j|j ddddf} tj||j|j ddddf} || |k\| |kz| |k\z| |kzzS)a Crop masks to bounding box regions. Args: masks (torch.Tensor): Masks with shape (N, H, W). boxes (torch.Tensor): Bounding box coordinates with shape (N, 4) in relative point form. Returns: (torch.Tensor): Cropped masks. Nr1rCr)rgr!chunkrhr r0) rrM_rrx1y1x2y2rrs r crop_maskrskkGAq![[q!Tz!2Aq9NBB Qu||288rc|j\}}}||jj|dzjd||}t|d|d}t ||}|j dS)a Apply masks to bounding boxes using mask head output with native upsampling. Args: protos (torch.Tensor): Mask prototypes with shape (mask_dim, mask_h, mask_w). masks_in (torch.Tensor): Mask coefficients with shape (N, mask_dim) where N is number of masks after NMS. bboxes (torch.Tensor): Bounding boxes with shape (N, 4) where N is number of masks after NMS. shape (tuple): Input image size as (height, width). Returns: (torch.Tensor): Binary mask tensor with shape (H, W, N). raNrr$)rgr)ri scale_masksrr()r)r*r+rgrr,r-rs rprocess_mask_nativer6sn IAr2  ++Ar2 2 8 8R DE d U +A .E eV $E 99S>rc|jdd\}}t||dz ||dz }||d|zz ||d|zz g}|r|dxxdzcc<|dxxdzcc<|r4tt|ddz tt|ddz fnd\}}|tt|ddzz |tt|ddzz } } |d|| || f}t j ||dd }|S) aN Rescale segment masks to target shape. Args: masks (torch.Tensor): Masks with shape (N, C, H, W). shape (tuple): Target height and width as (height, width). padding (bool): Whether masks are based on YOLO-style augmented images with padding. Returns: (torch.Tensor): Rescaled masks. rDNrrCrE)rr.r"Fr#)rgr5rVrJr&r') rrgr@r,r-rPrQrrrrs rr5r5s[[_FB rE!H}b58m ,D a4 eAho!5 6C A!  A! HOU3q6C<()3uSVc\/B+CDU[IC Ss1v|$ %% Ss1v|$ %% F #s6z4:- .E MM%Zu ME Lr normalizecz|At|d|dz |d|dz }|d|d|zz dz |d|d|zz dz f}n |dd}|d}|r |dxx|dzcc<|dxx|dzcc<|dxx|zcc<|dxx|zcc<t||}|r |dxx|dzcc<|dxx|dzcc<|S)aT Rescale segment coordinates from img1_shape to img0_shape. Args: img1_shape (tuple): Shape of the source image. coords (torch.Tensor): Coordinates to scale with shape (N, 2). img0_shape (tuple): Shape of the target image. ratio_pad (tuple, optional): Ratio and padding values as ((ratio_h, ratio_w), (pad_h, pad_w)). normalize (bool): Whether to normalize coordinates to range [0, 1]. padding (bool): Whether coordinates are based on YOLO-style augmented images with padding. Returns: (torch.Tensor): Scaled coordinates. rrCrDrFrG)r5r)rLrrNrOr8r@rPrQs r scale_coordsr:s:a=:a=0*Q-*Q-2OP!}z!}t33q8:a=:VW=[_K_;_cd:dd|Alv#a& v#a&  6NdN 6NdN  ,Fv*Q-'v*Q-' Mrc@|jd\}}}}}|tjztjdz k\}tj|||}tj|||}|tjdz z}tj |||||gdS)z Regularize rotated bounding boxes to range [0, pi/2]. Args: rboxes (torch.Tensor): Input rotated boxes with shape (N, 5) in xywhr format. Returns: (torch.Tensor): Regularized rotated boxes. rarrD)unbindrWrr!rr) rr<r=rrrswapw_h_s rregularize_rboxesr@(sMMbM)MAq!Q tww;$''A+ %D T1a B T1a B TWWq[A ;;1b"a(b 11rstrategyc ddlm}g}|jjj j dD]8}t j|t jt jd}|r|dk(r]t|dkDr9tj||Dcgc]}|jddc}n|djdd}n{|dk(rvtj|tj|Dcgc] }t|c}jjdd}ntj d }|j#|j d ;|Scc}wcc}w) a& Convert masks to segments using contour detection. Args: masks (torch.Tensor): Binary masks with shape (batch_size, 160, 160). strategy (str): Segmentation strategy, either 'all' or 'largest'. Returns: (list): List of segment masks as float32 arrays. r)merge_multi_segmentuint8allrCrarDlargest)rrDr)ultralytics.data.converterrCrVrrastyper findContours RETR_EXTERNALCHAIN_APPROX_SIMPLErr3rrr4argmaxr:r)rrArCr r<rs rmasks2segmentsrM;s=?H YY[__  $ $ & - -g 6 -   Q 1 133J3J KA N 5 1vzNN#6RS7TQ "a8H7T#UV1b!, Y&HHQrxx(;AQ(;<CCEFGOOPRTUV A+, - O8U )<+]r)patternreplstring)resub)r s r clean_strr\gs 66>SQR SSrct|tjr%tj|tjSt j|t jS)zSCreate empty torch.Tensor or np.ndarray with same shape as input and float32 dtype.r/)rTr!rUrrr3)r<s rrrtsG5?q%,,4O%--0UWUbUbcdlnlvlvUwr)r^)NTF)?T)g?r_NFFr*i,rg?i0uiTFFF)N)r^r^rr)r^r^Fr$)i)F)T)NFT)rE)8 contextlibrWrZrtypingrrrr3r!torch.nn.functionalnn functionalr&ultralytics.utilsrultralytics.utils.metricsrContextDecoratorrrVr?r rRrZr)rurrKrrrrrrrrrrrr r rrr3r6r5r:r@r rMrUndarrayrUr\rr*rrrisH   $36#j))6#r3<#)#)[_#)L ,c , %$F  !R6R6R6  R6  R6R6 R6R6R6 R6R6R6R6 !R6j,(&R ( ( S 3 C 3 0 S 3 D u 4    d.+>  &"30A&DB(t:DcgD2&C> fU\\ fbjj f Tr