Commit ⋅ pytorch/vision

Showing 2 of 5 files from the diff.

80.84% 27.27% ø

torchvision/transforms/functional.py changed.

81.80% 100.00% ø

torchvision/transforms/transforms.py changed.

Other files ignored by Codecov

test/test_transforms_tensor.py has changed.

test/test_transforms.py has changed.

test/test_functional_tensor.py has changed.

@@ -20,7 +20,7 @@

Loading

from . import functional_tensor as F_t


class InterpolationModes(Enum):
class InterpolationMode(Enum):
    """Interpolation modes
    """
    NEAREST = "nearest"

@@ -33,26 +33,26 @@

Loading



# TODO: Once torchscript supports Enums with staticmethod
# this can be put into InterpolationModes as staticmethod
def _interpolation_modes_from_int(i: int) -> InterpolationModes:
# this can be put into InterpolationMode as staticmethod
def _interpolation_modes_from_int(i: int) -> InterpolationMode:
    inverse_modes_mapping = {
        0: InterpolationModes.NEAREST,
        2: InterpolationModes.BILINEAR,
        3: InterpolationModes.BICUBIC,
        4: InterpolationModes.BOX,
        5: InterpolationModes.HAMMING,
        1: InterpolationModes.LANCZOS,
        0: InterpolationMode.NEAREST,
        2: InterpolationMode.BILINEAR,
        3: InterpolationMode.BICUBIC,
        4: InterpolationMode.BOX,
        5: InterpolationMode.HAMMING,
        1: InterpolationMode.LANCZOS,
    }
    return inverse_modes_mapping[i]


pil_modes_mapping = {
    InterpolationModes.NEAREST: 0,
    InterpolationModes.BILINEAR: 2,
    InterpolationModes.BICUBIC: 3,
    InterpolationModes.BOX: 4,
    InterpolationModes.HAMMING: 5,
    InterpolationModes.LANCZOS: 1,
    InterpolationMode.NEAREST: 0,
    InterpolationMode.BILINEAR: 2,
    InterpolationMode.BICUBIC: 3,
    InterpolationMode.BOX: 4,
    InterpolationMode.HAMMING: 5,
    InterpolationMode.LANCZOS: 1,
}

_is_pil_image = F_pil._is_pil_image

@@ -329,7 +329,7 @@

Loading

    return tensor


def resize(img: Tensor, size: List[int], interpolation: InterpolationModes = InterpolationModes.BILINEAR) -> Tensor:
def resize(img: Tensor, size: List[int], interpolation: InterpolationMode = InterpolationMode.BILINEAR) -> Tensor:
    r"""Resize the input image to the given size.
    The image can be a PIL Image or a torch Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions

@@ -343,10 +343,10 @@

Loading

            :math:`\left(\text{size} \times \frac{\text{height}}{\text{width}}, \text{size}\right)`.
            In torchscript mode size as single int is not supported, use a tuple or
            list of length 1: ``[size, ]``.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`.
            Default is ``InterpolationModes.BILINEAR``. If input is Tensor, only ``InterpolationModes.NEAREST``,
            ``InterpolationModes.BILINEAR`` and ``InterpolationModes.BICUBIC`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`.
            Default is ``InterpolationMode.BILINEAR``. If input is Tensor, only ``InterpolationMode.NEAREST``,
            ``InterpolationMode.BILINEAR`` and ``InterpolationMode.BICUBIC`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.

    Returns:

@@ -355,13 +355,13 @@

Loading

    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
            "Argument interpolation should be of type InterpolationModes instead of int. "
            "Please, use InterpolationModes enum."
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
        )
        interpolation = _interpolation_modes_from_int(interpolation)

    if not isinstance(interpolation, InterpolationModes):
        raise TypeError("Argument interpolation should be a InterpolationModes")
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")

    if not isinstance(img, torch.Tensor):
        pil_interpolation = pil_modes_mapping[interpolation]

@@ -475,7 +475,7 @@

Loading


def resized_crop(
        img: Tensor, top: int, left: int, height: int, width: int, size: List[int],
        interpolation: InterpolationModes = InterpolationModes.BILINEAR
        interpolation: InterpolationMode = InterpolationMode.BILINEAR
) -> Tensor:
    """Crop the given image and resize it to desired size.
    The image can be a PIL Image or a Tensor, in which case it is expected

@@ -490,10 +490,10 @@

Loading

        height (int): Height of the crop box.
        width (int): Width of the crop box.
        size (sequence or int): Desired output size. Same semantics as ``resize``.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`.
            Default is ``InterpolationModes.BILINEAR``. If input is Tensor, only ``InterpolationModes.NEAREST``,
            ``InterpolationModes.BILINEAR`` and ``InterpolationModes.BICUBIC`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`.
            Default is ``InterpolationMode.BILINEAR``. If input is Tensor, only ``InterpolationMode.NEAREST``,
            ``InterpolationMode.BILINEAR`` and ``InterpolationMode.BICUBIC`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.

    Returns:

@@ -556,7 +556,7 @@

Loading

        img: Tensor,
        startpoints: List[List[int]],
        endpoints: List[List[int]],
        interpolation: InterpolationModes = InterpolationModes.BILINEAR,
        interpolation: InterpolationMode = InterpolationMode.BILINEAR,
        fill: Optional[int] = None
) -> Tensor:
    """Perform perspective transform of the given image.

@@ -569,9 +569,9 @@

Loading

            ``[top-left, top-right, bottom-right, bottom-left]`` of the original image.
        endpoints (list of list of ints): List containing four lists of two integers corresponding to four corners
            ``[top-left, top-right, bottom-right, bottom-left]`` of the transformed image.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
        fill (n-tuple or int or float): Pixel fill value for area outside the rotated
            image. If int or float, the value is used for all bands respectively.

@@ -587,13 +587,13 @@

Loading

    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
            "Argument interpolation should be of type InterpolationModes instead of int. "
            "Please, use InterpolationModes enum."
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
        )
        interpolation = _interpolation_modes_from_int(interpolation)

    if not isinstance(interpolation, InterpolationModes):
        raise TypeError("Argument interpolation should be a InterpolationModes")
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")

    if not isinstance(img, torch.Tensor):
        pil_interpolation = pil_modes_mapping[interpolation]

@@ -869,7 +869,7 @@

Loading



def rotate(
        img: Tensor, angle: float, interpolation: InterpolationModes = InterpolationModes.NEAREST,
        img: Tensor, angle: float, interpolation: InterpolationMode = InterpolationMode.NEAREST,
        expand: bool = False, center: Optional[List[int]] = None,
        fill: Optional[int] = None, resample: Optional[int] = None
) -> Tensor:

@@ -880,9 +880,9 @@

Loading

    Args:
        img (PIL Image or Tensor): image to be rotated.
        angle (float or int): rotation angle value in degrees, counter-clockwise.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
        expand (bool, optional): Optional expansion flag.
            If true, expands the output image to make it large enough to hold the entire rotated image.

@@ -913,8 +913,8 @@

Loading

    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
            "Argument interpolation should be of type InterpolationModes instead of int. "
            "Please, use InterpolationModes enum."
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
        )
        interpolation = _interpolation_modes_from_int(interpolation)


@@ -924,8 +924,8 @@

Loading

    if center is not None and not isinstance(center, (list, tuple)):
        raise TypeError("Argument center should be a sequence")

    if not isinstance(interpolation, InterpolationModes):
        raise TypeError("Argument interpolation should be a InterpolationModes")
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")

    if not isinstance(img, torch.Tensor):
        pil_interpolation = pil_modes_mapping[interpolation]

@@ -945,7 +945,7 @@

Loading


def affine(
        img: Tensor, angle: float, translate: List[int], scale: float, shear: List[float],
        interpolation: InterpolationModes = InterpolationModes.NEAREST, fill: Optional[int] = None,
        interpolation: InterpolationMode = InterpolationMode.NEAREST, fill: Optional[int] = None,
        resample: Optional[int] = None, fillcolor: Optional[int] = None
) -> Tensor:
    """Apply affine transformation on the image keeping image center invariant.

@@ -960,9 +960,9 @@

Loading

        shear (float or tuple or list): shear angle value in degrees between -180 to 180, clockwise direction.
            If a tuple of list is specified, the first value corresponds to a shear parallel to the x axis, while
            the second value corresponds to a shear parallel to the y axis.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
        fill (int): Optional fill color for the area outside the transform in the output image (Pillow>=5.0.0).
            This option is not supported for Tensor input. Fill value for the area outside the transform in the output

@@ -984,8 +984,8 @@

Loading

    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
            "Argument interpolation should be of type InterpolationModes instead of int. "
            "Please, use InterpolationModes enum."
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
        )
        interpolation = _interpolation_modes_from_int(interpolation)


@@ -1010,8 +1010,8 @@

Loading

    if not isinstance(shear, (numbers.Number, (list, tuple))):
        raise TypeError("Shear should be either a single value or a sequence of two values")

    if not isinstance(interpolation, InterpolationModes):
        raise TypeError("Argument interpolation should be a InterpolationModes")
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")

    if isinstance(angle, int):
        angle = float(angle)

@@ -14,14 +14,14 @@

Loading

    accimage = None

from . import functional as F
from .functional import InterpolationModes, _interpolation_modes_from_int
from .functional import InterpolationMode, _interpolation_modes_from_int


__all__ = ["Compose", "ToTensor", "PILToTensor", "ConvertImageDtype", "ToPILImage", "Normalize", "Resize", "Scale",
           "CenterCrop", "Pad", "Lambda", "RandomApply", "RandomChoice", "RandomOrder", "RandomCrop",
           "RandomHorizontalFlip", "RandomVerticalFlip", "RandomResizedCrop", "RandomSizedCrop", "FiveCrop", "TenCrop",
           "LinearTransformation", "ColorJitter", "RandomRotation", "RandomAffine", "Grayscale", "RandomGrayscale",
           "RandomPerspective", "RandomErasing", "GaussianBlur", "InterpolationModes"]
           "RandomPerspective", "RandomErasing", "GaussianBlur", "InterpolationMode"]


class Compose:

@@ -234,15 +234,15 @@

Loading

            (size * height / width, size).
            In torchscript mode padding as single int is not supported, use a tuple or
            list of length 1: ``[size, ]``.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` and
            ``InterpolationModes.BICUBIC`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` and
            ``InterpolationMode.BICUBIC`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.

    """

    def __init__(self, size, interpolation=InterpolationModes.BILINEAR):
    def __init__(self, size, interpolation=InterpolationMode.BILINEAR):
        super().__init__()
        if not isinstance(size, (int, Sequence)):
            raise TypeError("Size should be int or sequence. Got {}".format(type(size)))

@@ -253,8 +253,8 @@

Loading

        # Backward compatibility with integer value
        if isinstance(interpolation, int):
            warnings.warn(
                "Argument interpolation should be of type InterpolationModes instead of int. "
                "Please, use InterpolationModes enum."
                "Argument interpolation should be of type InterpolationMode instead of int. "
                "Please, use InterpolationMode enum."
            )
            interpolation = _interpolation_modes_from_int(interpolation)


@@ -663,9 +663,9 @@

Loading

        distortion_scale (float): argument to control the degree of distortion and ranges from 0 to 1.
            Default is 0.5.
        p (float): probability of the image being transformed. Default is 0.5.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
        fill (n-tuple or int or float): Pixel fill value for area outside the rotated
            image. If int or float, the value is used for all bands respectively. Default is 0.

@@ -673,15 +673,15 @@

Loading

            input. Fill value for the area outside the transform in the output image is always 0.
    """

    def __init__(self, distortion_scale=0.5, p=0.5, interpolation=InterpolationModes.BILINEAR, fill=0):
    def __init__(self, distortion_scale=0.5, p=0.5, interpolation=InterpolationMode.BILINEAR, fill=0):
        super().__init__()
        self.p = p

        # Backward compatibility with integer value
        if isinstance(interpolation, int):
            warnings.warn(
                "Argument interpolation should be of type InterpolationModes instead of int. "
                "Please, use InterpolationModes enum."
                "Argument interpolation should be of type InterpolationMode instead of int. "
                "Please, use InterpolationMode enum."
            )
            interpolation = _interpolation_modes_from_int(interpolation)


@@ -758,15 +758,15 @@

Loading

            made. If provided a tuple or list of length 1, it will be interpreted as (size[0], size[0]).
        scale (tuple of float): scale range of the cropped image before resizing, relatively to the origin image.
        ratio (tuple of float): aspect ratio range of the cropped image before resizing.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` and
            ``InterpolationModes.BICUBIC`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` and
            ``InterpolationMode.BICUBIC`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.

    """

    def __init__(self, size, scale=(0.08, 1.0), ratio=(3. / 4., 4. / 3.), interpolation=InterpolationModes.BILINEAR):
    def __init__(self, size, scale=(0.08, 1.0), ratio=(3. / 4., 4. / 3.), interpolation=InterpolationMode.BILINEAR):
        super().__init__()
        self.size = _setup_size(size, error_msg="Please provide only two dimensions (h, w) for size.")


@@ -780,8 +780,8 @@

Loading

        # Backward compatibility with integer value
        if isinstance(interpolation, int):
            warnings.warn(
                "Argument interpolation should be of type InterpolationModes instead of int. "
                "Please, use InterpolationModes enum."
                "Argument interpolation should be of type InterpolationMode instead of int. "
                "Please, use InterpolationMode enum."
            )
            interpolation = _interpolation_modes_from_int(interpolation)


@@ -1147,9 +1147,9 @@

Loading

        degrees (sequence or float or int): Range of degrees to select from.
            If degrees is a number instead of sequence like (min, max), the range of degrees
            will be (-degrees, +degrees).
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
        expand (bool, optional): Optional expansion flag.
            If true, expands the output to make it large enough to hold the entire rotated image.

@@ -1170,7 +1170,7 @@

Loading

    """

    def __init__(
        self, degrees, interpolation=InterpolationModes.NEAREST, expand=False, center=None, fill=None, resample=None
        self, degrees, interpolation=InterpolationMode.NEAREST, expand=False, center=None, fill=None, resample=None
    ):
        super().__init__()
        if resample is not None:

@@ -1182,8 +1182,8 @@

Loading

        # Backward compatibility with integer value
        if isinstance(interpolation, int):
            warnings.warn(
                "Argument interpolation should be of type InterpolationModes instead of int. "
                "Please, use InterpolationModes enum."
                "Argument interpolation should be of type InterpolationMode instead of int. "
                "Please, use InterpolationMode enum."
            )
            interpolation = _interpolation_modes_from_int(interpolation)


@@ -1253,9 +1253,9 @@

Loading

            range (shear[0], shear[1]) will be applied. Else if shear is a tuple or list of 4 values,
            a x-axis shear in (shear[0], shear[1]) and y-axis shear in (shear[2], shear[3]) will be applied.
            Will not apply shear by default.
        interpolation (InterpolationModes): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
            If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
        fill (tuple or int): Optional fill color (Tuple for RGB Image and int for grayscale) for the area
            outside the transform in the output image (Pillow>=5.0.0). This option is not supported for Tensor

@@ -1270,7 +1270,7 @@

Loading

    """

    def __init__(
        self, degrees, translate=None, scale=None, shear=None, interpolation=InterpolationModes.NEAREST, fill=0,
        self, degrees, translate=None, scale=None, shear=None, interpolation=InterpolationMode.NEAREST, fill=0,
        fillcolor=None, resample=None
    ):
        super().__init__()

@@ -1283,8 +1283,8 @@

Loading

        # Backward compatibility with integer value
        if isinstance(interpolation, int):
            warnings.warn(
                "Argument interpolation should be of type InterpolationModes instead of int. "
                "Please, use InterpolationModes enum."
                "Argument interpolation should be of type InterpolationMode instead of int. "
                "Please, use InterpolationMode enum."
            )
            interpolation = _interpolation_modes_from_int(interpolation)


@@ -1377,7 +1377,7 @@

Loading

            s += ', scale={scale}'
        if self.shear is not None:
            s += ', shear={shear}'
        if self.interpolation != InterpolationModes.NEAREST:
        if self.interpolation != InterpolationMode.NEAREST:
            s += ', interpolation={interpolation}'
        if self.fill != 0:
            s += ', fill={fill}'

Files		•	•	•	Coverage
torchvision	8,899	6,483	445	1,971	72.85%
Project Totals (99 files)	8,899	6,483	445	1,971	72.85%

View CI Build Download

No yaml found.

Create your codecov.yml to customize your Codecov experience

Read the docs

Interact now

The inner-most circle is the entire project, moving away from the center are folders then, finally, a single file. The size and color of each slice is representing the number of statements and the coverage, respectively.

SVG

The top section represents the entire project. Proceeding with folders and finally individual files. The size and color of each slice is representing the number of statements and the coverage, respectively.

SVG

Each block represents a single file in the project. The size and color of each block is represented by the number of statements and the coverage, respectively.

SVG

Learn more about the coverage sunburst graph

329	329		return tensor
330	330
331	331
332		-	def resize(img: Tensor, size: List[int], interpolation: InterpolationModes = InterpolationModes.BILINEAR) -> Tensor:
	332	+	def resize(img: Tensor, size: List[int], interpolation: InterpolationMode = InterpolationMode.BILINEAR) -> Tensor:
333	333		r"""Resize the input image to the given size.
334	334		The image can be a PIL Image or a torch Tensor, in which case it is expected
335	335		to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions

343	343		:math:`\left(\text{size} \times \frac{\text{height}}{\text{width}}, \text{size}\right)`.
344	344		In torchscript mode size as single int is not supported, use a tuple or
345	345		list of length 1: ``[size, ]``.
346		-	interpolation (InterpolationModes): Desired interpolation enum defined by
347		-	:class:`torchvision.transforms.InterpolationModes`.
348		-	Default is ``InterpolationModes.BILINEAR``. If input is Tensor, only ``InterpolationModes.NEAREST``,
349		-	``InterpolationModes.BILINEAR`` and ``InterpolationModes.BICUBIC`` are supported.
	346	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	347	+	:class:`torchvision.transforms.InterpolationMode`.
	348	+	Default is ``InterpolationMode.BILINEAR``. If input is Tensor, only ``InterpolationMode.NEAREST``,
	349	+	``InterpolationMode.BILINEAR`` and ``InterpolationMode.BICUBIC`` are supported.
350	350		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
351	351
352	352		Returns:

355	355		# Backward compatibility with integer value
356	356		if isinstance(interpolation, int):
357	357		warnings.warn(
358		-	"Argument interpolation should be of type InterpolationModes instead of int. "
359		-	"Please, use InterpolationModes enum."
	358	+	"Argument interpolation should be of type InterpolationMode instead of int. "
	359	+	"Please, use InterpolationMode enum."
360	360		)
361	361		interpolation = _interpolation_modes_from_int(interpolation)
362	362
363		-	if not isinstance(interpolation, InterpolationModes):
364		-	raise TypeError("Argument interpolation should be a InterpolationModes")
	363	+	if not isinstance(interpolation, InterpolationMode):
	364	+	raise TypeError("Argument interpolation should be a InterpolationMode")
365	365
366	366		if not isinstance(img, torch.Tensor):
367	367		pil_interpolation = pil_modes_mapping[interpolation]

475	475
476	476		def resized_crop(
477	477		img: Tensor, top: int, left: int, height: int, width: int, size: List[int],
478		-	interpolation: InterpolationModes = InterpolationModes.BILINEAR
	478	+	interpolation: InterpolationMode = InterpolationMode.BILINEAR
479	479		) -> Tensor:
480	480		"""Crop the given image and resize it to desired size.
481	481		The image can be a PIL Image or a Tensor, in which case it is expected

490	490		height (int): Height of the crop box.
491	491		width (int): Width of the crop box.
492	492		size (sequence or int): Desired output size. Same semantics as ``resize``.
493		-	interpolation (InterpolationModes): Desired interpolation enum defined by
494		-	:class:`torchvision.transforms.InterpolationModes`.
495		-	Default is ``InterpolationModes.BILINEAR``. If input is Tensor, only ``InterpolationModes.NEAREST``,
496		-	``InterpolationModes.BILINEAR`` and ``InterpolationModes.BICUBIC`` are supported.
	493	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	494	+	:class:`torchvision.transforms.InterpolationMode`.
	495	+	Default is ``InterpolationMode.BILINEAR``. If input is Tensor, only ``InterpolationMode.NEAREST``,
	496	+	``InterpolationMode.BILINEAR`` and ``InterpolationMode.BICUBIC`` are supported.
497	497		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
498	498
499	499		Returns:

20	20		from . import functional_tensor as F_t
21	21
22	22
23		-	class InterpolationModes(Enum):
	23	+	class InterpolationMode(Enum):
24	24		"""Interpolation modes
25	25		"""
26	26		NEAREST = "nearest"

33	33
34	34
35	35		# TODO: Once torchscript supports Enums with staticmethod
36		-	# this can be put into InterpolationModes as staticmethod
37		-	def _interpolation_modes_from_int(i: int) -> InterpolationModes:
	36	+	# this can be put into InterpolationMode as staticmethod
	37	+	def _interpolation_modes_from_int(i: int) -> InterpolationMode:
38	38		inverse_modes_mapping = {
39		-	0: InterpolationModes.NEAREST,
40		-	2: InterpolationModes.BILINEAR,
41		-	3: InterpolationModes.BICUBIC,
42		-	4: InterpolationModes.BOX,
43		-	5: InterpolationModes.HAMMING,
44		-	1: InterpolationModes.LANCZOS,
	39	+	0: InterpolationMode.NEAREST,
	40	+	2: InterpolationMode.BILINEAR,
	41	+	3: InterpolationMode.BICUBIC,
	42	+	4: InterpolationMode.BOX,
	43	+	5: InterpolationMode.HAMMING,
	44	+	1: InterpolationMode.LANCZOS,
45	45		}
46	46		return inverse_modes_mapping[i]
47	47
48	48
49	49		pil_modes_mapping = {
50		-	InterpolationModes.NEAREST: 0,
51		-	InterpolationModes.BILINEAR: 2,
52		-	InterpolationModes.BICUBIC: 3,
53		-	InterpolationModes.BOX: 4,
54		-	InterpolationModes.HAMMING: 5,
55		-	InterpolationModes.LANCZOS: 1,
	50	+	InterpolationMode.NEAREST: 0,
	51	+	InterpolationMode.BILINEAR: 2,
	52	+	InterpolationMode.BICUBIC: 3,
	53	+	InterpolationMode.BOX: 4,
	54	+	InterpolationMode.HAMMING: 5,
	55	+	InterpolationMode.LANCZOS: 1,
56	56		}
57	57
58	58		_is_pil_image = F_pil._is_pil_image

556	556		img: Tensor,
557	557		startpoints: List[List[int]],
558	558		endpoints: List[List[int]],
559		-	interpolation: InterpolationModes = InterpolationModes.BILINEAR,
	559	+	interpolation: InterpolationMode = InterpolationMode.BILINEAR,
560	560		fill: Optional[int] = None
561	561		) -> Tensor:
562	562		"""Perform perspective transform of the given image.

569	569		``[top-left, top-right, bottom-right, bottom-left]`` of the original image.
570	570		endpoints (list of list of ints): List containing four lists of two integers corresponding to four corners
571	571		``[top-left, top-right, bottom-right, bottom-left]`` of the transformed image.
572		-	interpolation (InterpolationModes): Desired interpolation enum defined by
573		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
574		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
	572	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	573	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
	574	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
575	575		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
576	576		fill (n-tuple or int or float): Pixel fill value for area outside the rotated
577	577		image. If int or float, the value is used for all bands respectively.

869	869
870	870
871	871		def rotate(
872		-	img: Tensor, angle: float, interpolation: InterpolationModes = InterpolationModes.NEAREST,
	872	+	img: Tensor, angle: float, interpolation: InterpolationMode = InterpolationMode.NEAREST,
873	873		expand: bool = False, center: Optional[List[int]] = None,
874	874		fill: Optional[int] = None, resample: Optional[int] = None
875	875		) -> Tensor:

880	880		Args:
881	881		img (PIL Image or Tensor): image to be rotated.
882	882		angle (float or int): rotation angle value in degrees, counter-clockwise.
883		-	interpolation (InterpolationModes): Desired interpolation enum defined by
884		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
885		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
	883	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	884	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
	885	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
886	886		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
887	887		expand (bool, optional): Optional expansion flag.
888	888		If true, expands the output image to make it large enough to hold the entire rotated image.

924	924		if center is not None and not isinstance(center, (list, tuple)):
925	925		raise TypeError("Argument center should be a sequence")
926	926
927		-	if not isinstance(interpolation, InterpolationModes):
928		-	raise TypeError("Argument interpolation should be a InterpolationModes")
	927	+	if not isinstance(interpolation, InterpolationMode):
	928	+	raise TypeError("Argument interpolation should be a InterpolationMode")
929	929
930	930		if not isinstance(img, torch.Tensor):
931	931		pil_interpolation = pil_modes_mapping[interpolation]

945	945
946	946		def affine(
947	947		img: Tensor, angle: float, translate: List[int], scale: float, shear: List[float],
948		-	interpolation: InterpolationModes = InterpolationModes.NEAREST, fill: Optional[int] = None,
	948	+	interpolation: InterpolationMode = InterpolationMode.NEAREST, fill: Optional[int] = None,
949	949		resample: Optional[int] = None, fillcolor: Optional[int] = None
950	950		) -> Tensor:
951	951		"""Apply affine transformation on the image keeping image center invariant.

960	960		shear (float or tuple or list): shear angle value in degrees between -180 to 180, clockwise direction.
961	961		If a tuple of list is specified, the first value corresponds to a shear parallel to the x axis, while
962	962		the second value corresponds to a shear parallel to the y axis.
963		-	interpolation (InterpolationModes): Desired interpolation enum defined by
964		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
965		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
	963	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	964	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
	965	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
966	966		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
967	967		fill (int): Optional fill color for the area outside the transform in the output image (Pillow>=5.0.0).
968	968		This option is not supported for Tensor input. Fill value for the area outside the transform in the output

1010	1010		if not isinstance(shear, (numbers.Number, (list, tuple))):
1011	1011		raise TypeError("Shear should be either a single value or a sequence of two values")
1012	1012
1013		-	if not isinstance(interpolation, InterpolationModes):
1014		-	raise TypeError("Argument interpolation should be a InterpolationModes")
	1013	+	if not isinstance(interpolation, InterpolationMode):
	1014	+	raise TypeError("Argument interpolation should be a InterpolationMode")
1015	1015
1016	1016		if isinstance(angle, int):
1017	1017		angle = float(angle)

1377	1377		s += ', scale={scale}'
1378	1378		if self.shear is not None:
1379	1379		s += ', shear={shear}'
1380		-	if self.interpolation != InterpolationModes.NEAREST:
	1380	+	if self.interpolation != InterpolationMode.NEAREST:
1381	1381		s += ', interpolation={interpolation}'
1382	1382		if self.fill != 0:
1383	1383		s += ', fill={fill}'

Navigation	Overlay
`t` Navigate files	`h` Toggle hits
`y` Change url to tip of branch	`m` Toggle misses
`b / v` Jump to prev/next hit line	`p` Toggle partial
`z / x` Jump to prev/next missed or partial line	`1..9` Toggle flags
`shift + o` Open current page in GitHub	`a` Toggle all on
`/ or ?` Show keyboard shortcuts dialog	`c` Toggle context lines or commits

14	14		accimage = None
15	15
16	16		from . import functional as F
17		-	from .functional import InterpolationModes, _interpolation_modes_from_int
	17	+	from .functional import InterpolationMode, _interpolation_modes_from_int
18	18
19	19
20	20		__all__ = ["Compose", "ToTensor", "PILToTensor", "ConvertImageDtype", "ToPILImage", "Normalize", "Resize", "Scale",
21	21		"CenterCrop", "Pad", "Lambda", "RandomApply", "RandomChoice", "RandomOrder", "RandomCrop",
22	22		"RandomHorizontalFlip", "RandomVerticalFlip", "RandomResizedCrop", "RandomSizedCrop", "FiveCrop", "TenCrop",
23	23		"LinearTransformation", "ColorJitter", "RandomRotation", "RandomAffine", "Grayscale", "RandomGrayscale",
24		-	"RandomPerspective", "RandomErasing", "GaussianBlur", "InterpolationModes"]
	24	+	"RandomPerspective", "RandomErasing", "GaussianBlur", "InterpolationMode"]
25	25
26	26
27	27		class Compose:

234	234		(size * height / width, size).
235	235		In torchscript mode padding as single int is not supported, use a tuple or
236	236		list of length 1: ``[size, ]``.
237		-	interpolation (InterpolationModes): Desired interpolation enum defined by
238		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
239		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` and
240		-	``InterpolationModes.BICUBIC`` are supported.
	237	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	238	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
	239	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` and
	240	+	``InterpolationMode.BICUBIC`` are supported.
241	241		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
242	242
243	243		"""
244	244
245		-	def __init__(self, size, interpolation=InterpolationModes.BILINEAR):
	245	+	def __init__(self, size, interpolation=InterpolationMode.BILINEAR):
246	246		super().__init__()
247	247		if not isinstance(size, (int, Sequence)):
248	248		raise TypeError("Size should be int or sequence. Got {}".format(type(size)))

663	663		distortion_scale (float): argument to control the degree of distortion and ranges from 0 to 1.
664	664		Default is 0.5.
665	665		p (float): probability of the image being transformed. Default is 0.5.
666		-	interpolation (InterpolationModes): Desired interpolation enum defined by
667		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
668		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
	666	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	667	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
	668	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
669	669		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
670	670		fill (n-tuple or int or float): Pixel fill value for area outside the rotated
671	671		image. If int or float, the value is used for all bands respectively. Default is 0.

673	673		input. Fill value for the area outside the transform in the output image is always 0.
674	674		"""
675	675
676		-	def __init__(self, distortion_scale=0.5, p=0.5, interpolation=InterpolationModes.BILINEAR, fill=0):
	676	+	def __init__(self, distortion_scale=0.5, p=0.5, interpolation=InterpolationMode.BILINEAR, fill=0):
677	677		super().__init__()
678	678		self.p = p
679	679
680	680		# Backward compatibility with integer value
681	681		if isinstance(interpolation, int):
682	682		warnings.warn(
683		-	"Argument interpolation should be of type InterpolationModes instead of int. "
684		-	"Please, use InterpolationModes enum."
	683	+	"Argument interpolation should be of type InterpolationMode instead of int. "
	684	+	"Please, use InterpolationMode enum."
685	685		)
686	686		interpolation = _interpolation_modes_from_int(interpolation)
687	687

758	758		made. If provided a tuple or list of length 1, it will be interpreted as (size[0], size[0]).
759	759		scale (tuple of float): scale range of the cropped image before resizing, relatively to the origin image.
760	760		ratio (tuple of float): aspect ratio range of the cropped image before resizing.
761		-	interpolation (InterpolationModes): Desired interpolation enum defined by
762		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.BILINEAR``.
763		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` and
764		-	``InterpolationModes.BICUBIC`` are supported.
	761	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	762	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
	763	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` and
	764	+	``InterpolationMode.BICUBIC`` are supported.
765	765		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
766	766
767	767		"""
768	768
769		-	def __init__(self, size, scale=(0.08, 1.0), ratio=(3. / 4., 4. / 3.), interpolation=InterpolationModes.BILINEAR):
	769	+	def __init__(self, size, scale=(0.08, 1.0), ratio=(3. / 4., 4. / 3.), interpolation=InterpolationMode.BILINEAR):
770	770		super().__init__()
771	771		self.size = _setup_size(size, error_msg="Please provide only two dimensions (h, w) for size.")
772	772

1147	1147		degrees (sequence or float or int): Range of degrees to select from.
1148	1148		If degrees is a number instead of sequence like (min, max), the range of degrees
1149	1149		will be (-degrees, +degrees).
1150		-	interpolation (InterpolationModes): Desired interpolation enum defined by
1151		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
1152		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
	1150	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	1151	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
	1152	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
1153	1153		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
1154	1154		expand (bool, optional): Optional expansion flag.
1155	1155		If true, expands the output to make it large enough to hold the entire rotated image.

1170	1170		"""
1171	1171
1172	1172		def __init__(
1173		-	self, degrees, interpolation=InterpolationModes.NEAREST, expand=False, center=None, fill=None, resample=None
	1173	+	self, degrees, interpolation=InterpolationMode.NEAREST, expand=False, center=None, fill=None, resample=None
1174	1174		):
1175	1175		super().__init__()
1176	1176		if resample is not None:

1182	1182		# Backward compatibility with integer value
1183	1183		if isinstance(interpolation, int):
1184	1184		warnings.warn(
1185		-	"Argument interpolation should be of type InterpolationModes instead of int. "
1186		-	"Please, use InterpolationModes enum."
	1185	+	"Argument interpolation should be of type InterpolationMode instead of int. "
	1186	+	"Please, use InterpolationMode enum."
1187	1187		)
1188	1188		interpolation = _interpolation_modes_from_int(interpolation)
1189	1189

1253	1253		range (shear[0], shear[1]) will be applied. Else if shear is a tuple or list of 4 values,
1254	1254		a x-axis shear in (shear[0], shear[1]) and y-axis shear in (shear[2], shear[3]) will be applied.
1255	1255		Will not apply shear by default.
1256		-	interpolation (InterpolationModes): Desired interpolation enum defined by
1257		-	:class:`torchvision.transforms.InterpolationModes`. Default is ``InterpolationModes.NEAREST``.
1258		-	If input is Tensor, only ``InterpolationModes.NEAREST``, ``InterpolationModes.BILINEAR`` are supported.
	1256	+	interpolation (InterpolationMode): Desired interpolation enum defined by
	1257	+	:class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
	1258	+	If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
1259	1259		For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
1260	1260		fill (tuple or int): Optional fill color (Tuple for RGB Image and int for grayscale) for the area
1261	1261		outside the transform in the output image (Pillow>=5.0.0). This option is not supported for Tensor

1270	1270		"""
1271	1271
1272	1272		def __init__(
1273		-	self, degrees, translate=None, scale=None, shear=None, interpolation=InterpolationModes.NEAREST, fill=0,
	1273	+	self, degrees, translate=None, scale=None, shear=None, interpolation=InterpolationMode.NEAREST, fill=0,
1274	1274		fillcolor=None, resample=None
1275	1275		):
1276	1276		super().__init__()

1283	1283		# Backward compatibility with integer value
1284	1284		if isinstance(interpolation, int):
1285	1285		warnings.warn(
1286		-	"Argument interpolation should be of type InterpolationModes instead of int. "
1287		-	"Please, use InterpolationModes enum."
	1286	+	"Argument interpolation should be of type InterpolationMode instead of int. "
	1287	+	"Please, use InterpolationMode enum."
1288	1288		)
1289	1289		interpolation = _interpolation_modes_from_int(interpolation)
1290	1290