3jRSSKJrJr SSKrSSKJr "SS\R 5rg))AnyCallableN)nnc^\rSrSrSrS\S\R4SS4U4SjjrS\RS \ S \ S\R4S jr S r U=r $) Lambdaa8Applies user-defined lambda as a transform. Args: func: Callable function. Returns: The output of the user-defined lambda. Example: >>> import kornia >>> x = torch.rand(1, 3, 5, 5) >>> f = Lambda(lambda x: kornia.color.rgb_to_grayscale(x)) >>> f(x).shape torch.Size([1, 1, 5, 5]) func.returnNc>[TU]5 [U5(d"[S[ U5R <35eXlg)Nz'Argument lambd should be callable, got )super__init__callable TypeErrortype__name__r )selfr __class__s V/home/wildlama/miniconda3/lib/python3.13/site-packages/kornia/contrib/lambda_module.pyr Lambda.__init__*s: ~~Ed4jFYFYE\]^ ^ imgargskwargsc.UR"U/UQ70UD6$)aRun the user-provided transform function. This method simply forwards all arguments to ``self.func``. It is useful when you want to insert a custom Python callable inside a ``torch.nn.Module`` pipeline. Args: img: Primary input tensor passed as the first positional argument. For image tasks this is commonly shaped ``(B, C, H, W)``, where ``B`` = batch size, ``C`` = number of channels, ``H`` = height, and ``W`` = width. *args: Additional positional arguments forwarded to ``self.func``. **kwargs: Additional keyword arguments forwarded to ``self.func``. Returns: The output tensor returned by ``self.func``. r )rrrrs rforwardLambda.forward1s$yy.t.v..rr) r __module__ __qualname____firstlineno____doc__rtorchTensorr rr__static_attributes__ __classcell__)rs@rrrsW"Xc5<<&78T/5<<//s/u||//rr)typingrrr"rModulerrrr)s $! +/RYY+/r