xhZddlmZddlmZmZmZddlZddlmZddl m Z Gdde Z y) ) annotations)AnyMappingSequenceN) tree_flatten)TVTensorceZdZUdZded<eddd dZdddd dd Ze d dd Zdd dd Z y) KeyPointsu! :class:`torch.Tensor` subclass for tensors with shape ``[..., 2]`` that represent points in an image. .. note:: Support for keypoints was released in TorchVision 0.23 and is currently a BETA feature. We don't expect the API to change, but there may be some rare edge-cases. If you find any issues, please report them on our bug tracker: https://github.com/pytorch/vision/issues?q=is:open+is:issue Each point is represented by its X and Y coordinates along the width and height dimensions, respectively. Each point is represented by its X and Y coordinates along the width and height dimensions, respectively. KeyPoints may represent any object that can be represented by sequences of 2D points: - `Polygonal chains `_, including polylines, Bézier curves, etc., which can be of shape ``[N_chains, N_points, 2]``. - Polygons, which can be of shape ``[N_polygons, N_points, 2]``. - Skeletons, which can be of shape ``[N_skeletons, N_bones, 2, 2]`` for pose-estimation models. .. note:: Like for :class:`torchvision.tv_tensors.BoundingBoxes`, there should only be a single instance of the :class:`torchvision.tv_tensors.KeyPoints` class per sample e.g. ``{"img": img, "poins_of_interest": KeyPoints(...)}``, although one :class:`torchvision.tv_tensors.KeyPoints` object can contain multiple key points Args: data: Any data that can be turned into a tensor with :func:`torch.as_tensor`. canvas_size (two-tuple of ints): Height and width of the corresponding image or video. dtype (torch.dtype, optional): Desired data type of the bounding box. If omitted, will be inferred from ``data``. device (torch.device, optional): Desired device of the bounding box. If omitted and ``data`` is a :class:`torch.Tensor`, the device is taken from it. Otherwise, the bounding box is constructed on the CPU. requires_grad (bool, optional): Whether autograd should record operations on the bounding box. If omitted and ``data`` is a :class:`torch.Tensor`, the value is taken from it. Otherwise, defaults to ``False``. tuple[int, int] canvas_sizeT) check_dimsc|rK|jdk(r|jd}n*|jddk7rtd|j|j |}||_|S)Nrrz)Expected a tensor of shape (..., 2), not )ndim unsqueezeshape ValueError as_subclassr )clstensorr rpointss }/home/developers/rajanand/mypropertyqr-fmb-refixing-v2/venv/lib/python3.12/site-packages/torchvision/tv_tensors/_keypoints.py_wrapzKeyPoints._wrap;sf {{a))!,b!Q& #LV\\N![\\##C(( Ndtypedevice requires_gradcR|j||||}|j||S)Nrr ) _to_tensorr)rdatar rrr rs r__new__zKeyPoints.__new__Fs.E&P]^yy[y99rct||rt|jndz\}}td|D}|jt |t jr*t |tstj|d}|St |ttfrt|fd|D}|S)Nc3BK|]}t|ts|yw)N) isinstancer ).0xs r z)KeyPoints._wrap_output..[s(\q:aQZC[(\sFr rc3NK|]}tj|dyw)Fr-N)r r)r*partr s rr,z)KeyPoints._wrap_output..bs"!vgk)//$K\a/"b!vs"%) rtuplevaluesnextr r)torchTensorr rlisttype)routputargskwargs flat_params_first_keypoints_from_argsr s @r _wrap_outputzKeyPoints._wrap_outputRs&deFMMO.DTV&WX Q$((\K(\$\!/;; fell +Jvy4Q__VQV_WF  .!T&\!vou!vvF r)tensor_contentsc:|j|jS)Nr") _make_reprr )selfr>s r__repr__zKeyPoints.__repr__es4+;+;<rrEstr) __name__ __module__ __qualname____doc____annotations__ classmethodrr%r=rBr'rrr r s+Z! ]a%)26%) : :% : " : 0 :# :  :!+/ )   $26=rr ) __future__rtypingrrrr3torch.utils._pytreer _tv_tensorr r r'rrrQs%")) , [=[=r