hESrSSKJr SSKJrJr SSKrSSKrSSK J r SSK J r J r JrJr SS/r"S S\ 5r"S S\ 5rg) aModule containing PyTorch-specific transforms for Albumentations. This module provides transforms that convert NumPy arrays to PyTorch tensors in the appropriate format. It handles both 2D image data and 3D volumetric data, ensuring that the tensor dimensions are correctly arranged according to PyTorch's expected format (channels first). These transforms are typically used as the final step in an augmentation pipeline before feeding data to a PyTorch model. ) annotations)AnyoverloadN)BasicTransform)MONO_CHANNEL_DIMENSIONSNUM_MULTI_CHANNEL_DIMENSIONSNUM_VOLUME_DIMENSIONSTargets ToTensor3D ToTensorV2c^\rSrSrSr\R \R4rS S U4Sjjjr \ SSj5r SSjr SSjr \SSj5r\SSj5rSS jrSS jrS rU=r$)r aConverts images/masks to PyTorch Tensors, inheriting from BasicTransform. For images: - If input is in `HWC` format, converts to PyTorch `CHW` format - If input is in `HW` format, converts to PyTorch `1HW` format (adds channel dimension) Attributes: transpose_mask (bool): If True, transposes 3D input mask dimensions from `[height, width, num_channels]` to `[num_channels, height, width]`. p (float): Probability of applying the transform. Default: 1.0. c,>[TU]US9 XlgN)p)super__init__transpose_mask)selfrr __class__s [/var/www/fran/franai/venv/lib/python3.13/site-packages/albumentations/pytorch/transforms.pyrToTensorV2.__init__+s 1,c`URURURURS.$)zDefine mapping of target name to target function. Returns: dict[str, Any]: Dictionary mapping target names to corresponding transform functions. )imageimagesmaskmasks)applyapply_to_images apply_to_maskapply_to_masksrs rtargetsToTensorV2.targets/s0ZZ**&&((   rc UR[[1;a Sn[U5eUR[:Xa[R "US5n[ R"URSSS55$)aConvert a 2D image array to a PyTorch tensor. Converts image from HWC or HW format to CHW format, handling both single-channel and multi-channel images. Args: img (np.ndarray): Image as a numpy array of shape (H,W) or (H,W,C) **params (Any): Additional parameters Returns: torch.Tensor: PyTorch tensor in CHW format Raises: ValueError: If image dimensions are neither HW nor HWC z7Albumentations only supports images in HW or HWC formatr) ndimrr ValueErrornp expand_dimstorch from_numpy transpose)rimgparamsmsgs rrToTensorV2.apply>sa" 8835QR RKCS/ ! 88. ...a(C aA 677rc UR(a'UR[:XaURSSS5n[R "U5$)a*Convert a mask array to a PyTorch tensor. If transpose_mask is True and mask has 3 dimensions (H,W,C), converts mask to channels-first format (C,H,W). Args: mask (np.ndarray): Mask as a numpy array **params (Any): Additional parameters Returns: torch.Tensor: PyTorch tensor of mask r'rr()rr)rr/r-r.)rrr1s rr!ToTensorV2.apply_to_maskXs<   4990L#L>>!Q*D%%rc gNrrr1s rr"ToTensorV2.apply_to_masksjs\_rc gr7r8r9s rr"r:msPSrc  [U[5(a#UVs/sHo0R"U40UD6PM sn$UR(a+UR[ :Xa[ R"US5n[R"U5$s snf)aMConvert numpy array or list of numpy array masks to torch tensor(s). Args: masks (np.ndarray | list[np.ndarray]): Numpy array of shape (N, H, W) or (N, H, W, C), or a list of numpy arrays with shape (H, W) or (H, W, C). **params (Any): Additional parameters. Returns: torch.Tensor | list[torch.Tensor]: If transpose_mask is True and input is (N, H, W, C), returns tensor of shape (N, C, H, W). If transpose_mask is True and input is (H, W, C), r eturns a list of tensors with shape (C, H, W). Otherwise, returns tensors with the same shape as input. )rr(r') isinstancelistr!rr)r r+r/r-r.)rrr1rs rr"r:psp eT " "CHI54&&t6v65I I   5::1F#FLL 5E&& JsB c UR[:wa[SURS35e[R"UR SSSS55$)z:Convert batch of images from (N, H, W, C) to (N, C, H, W).z!Expected 4D array (N,H,W,C), got D arrayrr=r(r')r)r r*r-r.r/)rrr1s rr ToTensorV2.apply_to_imagessJ ;;/ /@ WUV V 0 0Aq! <==r)r)F?)rboolrfloatreturnzdict[str, Any])r0 np.ndarrayr1rrG torch.Tensor)rrHr1rrGrI)rzlist[np.ndarray]r1rrGzlist[torch.Tensor])rrHr1rrGrI)rznp.ndarray | list[np.ndarray]r1rrGz!torch.Tensor | list[torch.Tensor])rrHr1rrGrI)__name__ __module__ __qualname____firstlineno____doc__r IMAGEMASK_targetsrpropertyr$rr!rr"r __static_attributes__ __classcell__rs@rr r sq  w||,H--    84&$__ SS'*>>rc^\rSrSrSr\R \R4rSS U4Sjjjr \ S Sj5r S Sjr S Sjr SrU=r$) r aConvert 3D volumes and masks to PyTorch tensors. This transform is designed for 3D medical imaging data. It converts numpy arrays to PyTorch tensors and ensures consistent channel positioning. For all inputs (volumes and masks): - Input: (D, H, W, C) or (D, H, W) - depth, height, width, [channels] - Output: (C, D, H, W) - channels first format for PyTorch For single-channel input, adds C=1 dimension Note: This transform always moves channels to first position as this is the standard PyTorch format. For masks that need to stay in DHWC format, use a different transform or handle the transposition after this transform. Args: p (float): Probability of applying the transform. Default: 1.0 c >[TU]US9 gr)rr)rrrs rrToTensor3D.__init__s 1rc4URURS.$)zDefine mapping of target name to target function. Returns: dict[str, Any]: Dictionary mapping target names to corresponding transform functions )volumemask3d)apply_to_volumeapply_to_mask3dr#s rr$ToTensor3D.targetss ****  rc ,UR[:Xa([R"UR SSSS55$UR[S- :Xa)[R"U[ R S45$[SURS35e)z+Convert 3D volume to channels-first tensor.r=rr(r'.z2Expected 3D or 4D array (D,H,W) or (D,H,W,C), got rA)r)r r-r.r/r+newaxisr*)rr[r1s rr]ToTensor3D.apply_to_volumes{ ;;/ /##F$4$4Q1a$@A A ;;/!3 3##F2::s?$;< <Mfkk]Zabccrc (UR"U40UD6$)z)Convert 3D mask to channels-first tensor.)r])rr\r1s rr^ToTensor3D.apply_to_mask3ds##F5f55rr8)rC)rrErF)r[rHr1rrGrI)r\rHr1rrGrI)rJrKrLrMrNr rOrPrQrrRr$r]r^rSrTrUs@rr r sJ( w||,H    d66r)rN __future__rtypingrrnumpyr+r-(albumentations.core.transforms_interfacer$albumentations.core.type_definitionsrrr r __all__r r r8rrrksL#  C  &m>m>`1616r