torchvision.transforms¶
Transforms are common image transformations. They can be chained together using Compose
.
Additionally, there is the torchvision.transforms.functional
module.
Functional transforms give finegrained control over the transformations.
This is useful if you have to build a more complex transformation pipeline
(e.g. in the case of segmentation tasks).

class
torchvision.transforms.
Compose
(transforms)[source]¶ Composes several transforms together.
 Parameters
transforms (list of
Transform
objects) – list of transforms to compose.
Example
>>> transforms.Compose([ >>> transforms.CenterCrop(10), >>> transforms.ToTensor(), >>> ])
Transforms on PIL Image¶

class
torchvision.transforms.
CenterCrop
(size)[source]¶ Crops the given PIL Image at the center.
 Parameters
size (sequence or int) – Desired output size of the crop. If size is an int instead of sequence like (h, w), a square crop (size, size) is made.

class
torchvision.transforms.
ColorJitter
(brightness=0, contrast=0, saturation=0, hue=0)[source]¶ Randomly change the brightness, contrast and saturation of an image.
 Parameters
brightness (float or tuple of python:float (min, max)) – How much to jitter brightness. brightness_factor is chosen uniformly from [max(0, 1  brightness), 1 + brightness] or the given [min, max]. Should be non negative numbers.
contrast (float or tuple of python:float (min, max)) – How much to jitter contrast. contrast_factor is chosen uniformly from [max(0, 1  contrast), 1 + contrast] or the given [min, max]. Should be non negative numbers.
saturation (float or tuple of python:float (min, max)) – How much to jitter saturation. saturation_factor is chosen uniformly from [max(0, 1  saturation), 1 + saturation] or the given [min, max]. Should be non negative numbers.
hue (float or tuple of python:float (min, max)) – How much to jitter hue. hue_factor is chosen uniformly from [hue, hue] or the given [min, max]. Should have 0<= hue <= 0.5 or 0.5 <= min <= max <= 0.5.

class
torchvision.transforms.
FiveCrop
(size)[source]¶ Crop the given PIL Image into four corners and the central crop
Note
This transform returns a tuple of images and there may be a mismatch in the number of inputs and targets your Dataset returns. See below for an example of how to deal with this.
 Parameters
size (sequence or int) – Desired output size of the crop. If size is an
int
instead of sequence like (h, w), a square crop of size (size, size) is made.
Example
>>> transform = Compose([ >>> FiveCrop(size), # this is a list of PIL Images >>> Lambda(lambda crops: torch.stack([ToTensor()(crop) for crop in crops])) # returns a 4D tensor >>> ]) >>> #In your test loop you can do the following: >>> input, target = batch # input is a 5d tensor, target is 2d >>> bs, ncrops, c, h, w = input.size() >>> result = model(input.view(1, c, h, w)) # fuse batch size and ncrops >>> result_avg = result.view(bs, ncrops, 1).mean(1) # avg over crops

class
torchvision.transforms.
Grayscale
(num_output_channels=1)[source]¶ Convert image to grayscale.
 Parameters
num_output_channels (int) – (1 or 3) number of channels desired for output image
 Returns
Grayscale version of the input.  If num_output_channels == 1 : returned image is single channel  If num_output_channels == 3 : returned image is 3 channel with r == g == b
 Return type
PIL Image

class
torchvision.transforms.
Pad
(padding, fill=0, padding_mode='constant')[source]¶ Pad the given PIL Image on all sides with the given “pad” value.
 Parameters
padding (int or tuple) – Padding on each border. If a single int is provided this is used to pad all borders. If tuple of length 2 is provided this is the padding on left/right and top/bottom respectively. If a tuple of length 4 is provided this is the padding for the left, top, right and bottom borders respectively.
fill (int or tuple) – Pixel fill value for constant fill. Default is 0. If a tuple of length 3, it is used to fill R, G, B channels respectively. This value is only used when the padding_mode is constant
padding_mode (str) –
Type of padding. Should be: constant, edge, reflect or symmetric. Default is constant.
constant: pads with a constant value, this value is specified with fill
edge: pads with the last value at the edge of the image
reflect: pads with reflection of image without repeating the last value on the edge
For example, padding [1, 2, 3, 4] with 2 elements on both sides in reflect mode will result in [3, 2, 1, 2, 3, 4, 3, 2]
symmetric: pads with reflection of image repeating the last value on the edge
For example, padding [1, 2, 3, 4] with 2 elements on both sides in symmetric mode will result in [2, 1, 1, 2, 3, 4, 4, 3]

class
torchvision.transforms.
RandomAffine
(degrees, translate=None, scale=None, shear=None, resample=False, fillcolor=0)[source]¶ Random affine transformation of the image keeping center invariant
 Parameters
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). Set to 0 to deactivate rotations.
translate (tuple, optional) – tuple of maximum absolute fraction for horizontal and vertical translations. For example translate=(a, b), then horizontal shift is randomly sampled in the range img_width * a < dx < img_width * a and vertical shift is randomly sampled in the range img_height * b < dy < img_height * b. Will not translate by default.
scale (tuple, optional) – scaling factor interval, e.g (a, b), then scale is randomly sampled from the range a <= scale <= b. Will keep original scale by default.
shear (sequence or float or int, optional) – 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). Will not apply shear by default
resample ({PIL.Image.NEAREST, PIL.Image.BILINEAR, PIL.Image.BICUBIC}, optional) – An optional resampling filter. See filters for more information. If omitted, or if the image has mode “1” or “P”, it is set to PIL.Image.NEAREST.
fillcolor (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)

class
torchvision.transforms.
RandomApply
(transforms, p=0.5)[source]¶ Apply randomly a list of transformations with a given probability

class
torchvision.transforms.
RandomChoice
(transforms)[source]¶ Apply single transformation randomly picked from a list

class
torchvision.transforms.
RandomCrop
(size, padding=None, pad_if_needed=False, fill=0, padding_mode='constant')[source]¶ Crop the given PIL Image at a random location.
 Parameters
size (sequence or int) – Desired output size of the crop. If size is an int instead of sequence like (h, w), a square crop (size, size) is made.
padding (int or sequence, optional) – Optional padding on each border of the image. Default is None, i.e no padding. If a sequence of length 4 is provided, it is used to pad left, top, right, bottom borders respectively. If a sequence of length 2 is provided, it is used to pad left/right, top/bottom borders, respectively.
pad_if_needed (boolean) – It will pad the image if smaller than the desired size to avoid raising an exception. Since cropping is done after padding, the padding seems to be done at a random offset.
fill – Pixel fill value for constant fill. Default is 0. If a tuple of length 3, it is used to fill R, G, B channels respectively. This value is only used when the padding_mode is constant
padding_mode –
Type of padding. Should be: constant, edge, reflect or symmetric. Default is constant.
constant: pads with a constant value, this value is specified with fill
edge: pads with the last value on the edge of the image
reflect: pads with reflection of image (without repeating the last value on the edge)
padding [1, 2, 3, 4] with 2 elements on both sides in reflect mode will result in [3, 2, 1, 2, 3, 4, 3, 2]
symmetric: pads with reflection of image (repeating the last value on the edge)
padding [1, 2, 3, 4] with 2 elements on both sides in symmetric mode will result in [2, 1, 1, 2, 3, 4, 4, 3]

class
torchvision.transforms.
RandomGrayscale
(p=0.1)[source]¶ Randomly convert image to grayscale with a probability of p (default 0.1).
 Parameters
p (float) – probability that image should be converted to grayscale.
 Returns
Grayscale version of the input image with probability p and unchanged with probability (1p).  If input image is 1 channel: grayscale version is 1 channel  If input image is 3 channel: grayscale version is 3 channel with r == g == b
 Return type
PIL Image

class
torchvision.transforms.
RandomHorizontalFlip
(p=0.5)[source]¶ Horizontally flip the given PIL Image randomly with a given probability.
 Parameters
p (float) – probability of the image being flipped. Default value is 0.5

class
torchvision.transforms.
RandomOrder
(transforms)[source]¶ Apply a list of transformations in a random order

class
torchvision.transforms.
RandomPerspective
(distortion_scale=0.5, p=0.5, interpolation=3)[source]¶ Performs Perspective transformation of the given PIL Image randomly with a given probability.

class
torchvision.transforms.
RandomResizedCrop
(size, scale=(0.08, 1.0), ratio=(0.75, 1.3333333333333333), interpolation=2)[source]¶ Crop the given PIL Image to random size and aspect ratio.
A crop of random size (default: of 0.08 to 1.0) of the original size and a random aspect ratio (default: of 3/4 to 4/3) of the original aspect ratio is made. This crop is finally resized to given size. This is popularly used to train the Inception networks.
 Parameters
size – expected output size of each edge
scale – range of size of the origin size cropped
ratio – range of aspect ratio of the origin aspect ratio cropped
interpolation – Default: PIL.Image.BILINEAR

class
torchvision.transforms.
RandomRotation
(degrees, resample=False, expand=False, center=None)[source]¶ Rotate the image by angle.
 Parameters
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).
resample ({PIL.Image.NEAREST, PIL.Image.BILINEAR, PIL.Image.BICUBIC}, optional) – An optional resampling filter. See filters for more information. If omitted, or if the image has mode “1” or “P”, it is set to PIL.Image.NEAREST.
expand (bool, optional) – Optional expansion flag. If true, expands the output to make it large enough to hold the entire rotated image. If false or omitted, make the output image the same size as the input image. Note that the expand flag assumes rotation around the center and no translation.
center (2tuple, optional) – Optional center of rotation. Origin is the upper left corner. Default is the center of the image.

class
torchvision.transforms.
RandomSizedCrop
(*args, **kwargs)[source]¶ Note: This transform is deprecated in favor of RandomResizedCrop.

class
torchvision.transforms.
RandomVerticalFlip
(p=0.5)[source]¶ Vertically flip the given PIL Image randomly with a given probability.
 Parameters
p (float) – probability of the image being flipped. Default value is 0.5

class
torchvision.transforms.
Resize
(size, interpolation=2)[source]¶ Resize the input PIL Image to the given size.
 Parameters
size (sequence or int) – Desired output size. If size is a sequence like (h, w), output size will be matched to this. If size is an int, smaller edge of the image will be matched to this number. i.e, if height > width, then image will be rescaled to (size * height / width, size)
interpolation (int, optional) – Desired interpolation. Default is
PIL.Image.BILINEAR

class
torchvision.transforms.
Scale
(*args, **kwargs)[source]¶ Note: This transform is deprecated in favor of Resize.

class
torchvision.transforms.
TenCrop
(size, vertical_flip=False)[source]¶ Crop the given PIL Image into four corners and the central crop plus the flipped version of these (horizontal flipping is used by default)
Note
This transform returns a tuple of images and there may be a mismatch in the number of inputs and targets your Dataset returns. See below for an example of how to deal with this.
 Parameters
Example
>>> transform = Compose([ >>> TenCrop(size), # this is a list of PIL Images >>> Lambda(lambda crops: torch.stack([ToTensor()(crop) for crop in crops])) # returns a 4D tensor >>> ]) >>> #In your test loop you can do the following: >>> input, target = batch # input is a 5d tensor, target is 2d >>> bs, ncrops, c, h, w = input.size() >>> result = model(input.view(1, c, h, w)) # fuse batch size and ncrops >>> result_avg = result.view(bs, ncrops, 1).mean(1) # avg over crops
Transforms on torch.*Tensor¶

class
torchvision.transforms.
LinearTransformation
(transformation_matrix, mean_vector)[source]¶ Transform a tensor image with a square transformation matrix and a mean_vector computed offline. Given transformation_matrix and mean_vector, will flatten the torch.*Tensor and subtract mean_vector from it which is then followed by computing the dot product with the transformation matrix and then reshaping the tensor to its original shape.
 Applications:
whitening transformation: Suppose X is a column vector zerocentered data. Then compute the data covariance matrix [D x D] with torch.mm(X.t(), X), perform SVD on this matrix and pass it as transformation_matrix.

class
torchvision.transforms.
Normalize
(mean, std, inplace=False)[source]¶ Normalize a tensor image with mean and standard deviation. Given mean:
(M1,...,Mn)
and std:(S1,..,Sn)
forn
channels, this transform will normalize each channel of the inputtorch.*Tensor
i.e.input[channel] = (input[channel]  mean[channel]) / std[channel]
Note
This transform acts out of place, i.e., it does not mutates the input tensor.
 Parameters
mean (sequence) – Sequence of means for each channel.
std (sequence) – Sequence of standard deviations for each channel.
Conversion Transforms¶

class
torchvision.transforms.
ToPILImage
(mode=None)[source]¶ Convert a tensor or an ndarray to PIL Image.
Converts a torch.*Tensor of shape C x H x W or a numpy ndarray of shape H x W x C to a PIL Image while preserving the value range.
 Parameters
mode (PIL.Image mode) –
color space and pixel depth of input data (optional). If
mode
isNone
(default) there are some assumptions made about the input data:If the input has 4 channels, the
mode
is assumed to beRGBA
.If the input has 3 channels, the
mode
is assumed to beRGB
.If the input has 2 channels, the
mode
is assumed to beLA
.If the input has 1 channel, the
mode
is determined by the data type (i.eint
,float
,short
).

__call__
(pic)[source]¶  Parameters
pic (Tensor or numpy.ndarray) – Image to be converted to PIL Image.
 Returns
Image converted to PIL Image.
 Return type
PIL Image

class
torchvision.transforms.
ToTensor
[source]¶ Convert a
PIL Image
ornumpy.ndarray
to tensor.Converts a PIL Image or numpy.ndarray (H x W x C) in the range [0, 255] to a torch.FloatTensor of shape (C x H x W) in the range [0.0, 1.0] if the PIL Image belongs to one of the modes (L, LA, P, I, F, RGB, YCbCr, RGBA, CMYK, 1) or if the numpy.ndarray has dtype = np.uint8
In the other cases, tensors are returned without scaling.

__call__
(pic)[source]¶  Parameters
pic (PIL Image or numpy.ndarray) – Image to be converted to tensor.
 Returns
Converted image.
 Return type

Generic Transforms¶
Functional Transforms¶
Functional transforms give you finegrained control of the transformation pipeline. As opposed to the transformations above, functional transforms don’t contain a random number generator for their parameters. That means you have to specify/generate all parameters, but you can reuse the functional transform. For example, you can apply a functional transform to multiple images like this:
import torchvision.transforms.functional as TF
import random
def my_segmentation_transforms(image, segmentation):
if random.random() > 5:
angle = random.randint(30, 30)
image = TF.rotate(image, angle)
segmentation = TF.rotate(segmentation, angle)
# more transforms ...
return image, segmentation

torchvision.transforms.functional.
adjust_brightness
(img, brightness_factor)[source]¶ Adjust brightness of an Image.
 Parameters
img (PIL Image) – PIL Image to be adjusted.
brightness_factor (float) – How much to adjust the brightness. Can be any non negative number. 0 gives a black image, 1 gives the original image while 2 increases the brightness by a factor of 2.
 Returns
Brightness adjusted image.
 Return type
PIL Image

torchvision.transforms.functional.
adjust_contrast
(img, contrast_factor)[source]¶ Adjust contrast of an Image.
 Parameters
img (PIL Image) – PIL Image to be adjusted.
contrast_factor (float) – How much to adjust the contrast. Can be any non negative number. 0 gives a solid gray image, 1 gives the original image while 2 increases the contrast by a factor of 2.
 Returns
Contrast adjusted image.
 Return type
PIL Image

torchvision.transforms.functional.
adjust_gamma
(img, gamma, gain=1)[source]¶ Perform gamma correction on an image.
Also known as Power Law Transform. Intensities in RGB mode are adjusted based on the following equation:
\[I_{\text{out}} = 255 \times \text{gain} \times \left(\frac{I_{\text{in}}}{255}\right)^{\gamma} \]See Gamma Correction for more details.

torchvision.transforms.functional.
adjust_hue
(img, hue_factor)[source]¶ Adjust hue of an image.
The image hue is adjusted by converting the image to HSV and cyclically shifting the intensities in the hue channel (H). The image is then converted back to original image mode.
hue_factor is the amount of shift in H channel and must be in the interval [0.5, 0.5].
See Hue for more details.
 Parameters
img (PIL Image) – PIL Image to be adjusted.
hue_factor (float) – How much to shift the hue channel. Should be in [0.5, 0.5]. 0.5 and 0.5 give complete reversal of hue channel in HSV space in positive and negative direction respectively. 0 means no shift. Therefore, both 0.5 and 0.5 will give an image with complementary colors while 0 gives the original image.
 Returns
Hue adjusted image.
 Return type
PIL Image

torchvision.transforms.functional.
adjust_saturation
(img, saturation_factor)[source]¶ Adjust color saturation of an image.
 Parameters
img (PIL Image) – PIL Image to be adjusted.
saturation_factor (float) – How much to adjust the saturation. 0 will give a black and white image, 1 will give the original image while 2 will enhance the saturation by a factor of 2.
 Returns
Saturation adjusted image.
 Return type
PIL Image

torchvision.transforms.functional.
affine
(img, angle, translate, scale, shear, resample=0, fillcolor=None)[source]¶ Apply affine transformation on the image keeping image center invariant
 Parameters
img (PIL Image) – PIL Image to be rotated.
angle (float or int) – rotation angle in degrees between 180 and 180, clockwise direction.
translate (list or tuple of python:integers) – horizontal and vertical translations (postrotation translation)
scale (float) – overall scale
shear (float) – shear angle value in degrees between 180 to 180, clockwise direction.
resample (
PIL.Image.NEAREST
orPIL.Image.BILINEAR
orPIL.Image.BICUBIC
, optional) – An optional resampling filter. See filters for more information. If omitted, or if the image has mode “1” or “P”, it is set toPIL.Image.NEAREST
.fillcolor (int) – Optional fill color for the area outside the transform in the output image. (Pillow>=5.0.0)

torchvision.transforms.functional.
five_crop
(img, size)[source]¶ Crop the given PIL Image into four corners and the central crop.
Note
This transform returns a tuple of images and there may be a mismatch in the number of inputs and targets your
Dataset
returns.

torchvision.transforms.functional.
hflip
(img)[source]¶ Horizontally flip the given PIL Image.
 Parameters
img (PIL Image) – Image to be flipped.
 Returns
Horizontall flipped image.
 Return type
PIL Image

torchvision.transforms.functional.
normalize
(tensor, mean, std, inplace=False)[source]¶ Normalize a tensor image with mean and standard deviation.
Note
This transform acts out of place by default, i.e., it does not mutates the input tensor.
See
Normalize
for more details.

torchvision.transforms.functional.
pad
(img, padding, fill=0, padding_mode='constant')[source]¶ Pad the given PIL Image on all sides with specified padding mode and fill value.
 Parameters
img (PIL Image) – Image to be padded.
padding (int or tuple) – Padding on each border. If a single int is provided this is used to pad all borders. If tuple of length 2 is provided this is the padding on left/right and top/bottom respectively. If a tuple of length 4 is provided this is the padding for the left, top, right and bottom borders respectively.
fill – Pixel fill value for constant fill. Default is 0. If a tuple of length 3, it is used to fill R, G, B channels respectively. This value is only used when the padding_mode is constant
padding_mode –
Type of padding. Should be: constant, edge, reflect or symmetric. Default is constant.
constant: pads with a constant value, this value is specified with fill
edge: pads with the last value on the edge of the image
reflect: pads with reflection of image (without repeating the last value on the edge)
padding [1, 2, 3, 4] with 2 elements on both sides in reflect mode will result in [3, 2, 1, 2, 3, 4, 3, 2]
symmetric: pads with reflection of image (repeating the last value on the edge)
padding [1, 2, 3, 4] with 2 elements on both sides in symmetric mode will result in [2, 1, 1, 2, 3, 4, 4, 3]
 Returns
Padded image.
 Return type
PIL Image

torchvision.transforms.functional.
perspective
(img, startpoints, endpoints, interpolation=3)[source]¶ Perform perspective transform of the given PIL Image.
 Parameters
img (PIL Image) – Image to be transformed.
startpoints – List containing [topleft, topright, bottomright, bottomleft] of the orignal image
endpoints – List containing [topleft, topright, bottomright, bottomleft] of the transformed image
interpolation – Default Image.BICUBIC
 Returns
Perspectively transformed Image.
 Return type
PIL Image

torchvision.transforms.functional.
resize
(img, size, interpolation=2)[source]¶ Resize the input PIL Image to the given size.
 Parameters
img (PIL Image) – Image to be resized.
size (sequence or int) – Desired output size. If size is a sequence like (h, w), the output size will be matched to this. If size is an int, the smaller edge of the image will be matched to this number maintaing the aspect ratio. i.e, if height > width, then image will be rescaled to \(\left(\text{size} \times \frac{\text{height}}{\text{width}}, \text{size}\right)\)
interpolation (int, optional) – Desired interpolation. Default is
PIL.Image.BILINEAR
 Returns
Resized image.
 Return type
PIL Image

torchvision.transforms.functional.
resized_crop
(img, i, j, h, w, size, interpolation=2)[source]¶ Crop the given PIL Image and resize it to desired size.
Notably used in
RandomResizedCrop
. Parameters
img (PIL Image) – Image to be cropped.
i (int) – i in (i,j) i.e coordinates of the upper left corner
j (int) – j in (i,j) i.e coordinates of the upper left corner
h (int) – Height of the cropped image.
w (int) – Width of the cropped image.
size (sequence or int) – Desired output size. Same semantics as
resize
.interpolation (int, optional) – Desired interpolation. Default is
PIL.Image.BILINEAR
.
 Returns
Cropped image.
 Return type
PIL Image

torchvision.transforms.functional.
rotate
(img, angle, resample=False, expand=False, center=None)[source]¶ Rotate the image by angle.
 Parameters
img (PIL Image) – PIL Image to be rotated.
angle (float or int) – In degrees degrees counter clockwise order.
resample (
PIL.Image.NEAREST
orPIL.Image.BILINEAR
orPIL.Image.BICUBIC
, optional) – An optional resampling filter. See filters for more information. If omitted, or if the image has mode “1” or “P”, it is set toPIL.Image.NEAREST
.expand (bool, optional) – Optional expansion flag. If true, expands the output image to make it large enough to hold the entire rotated image. If false or omitted, make the output image the same size as the input image. Note that the expand flag assumes rotation around the center and no translation.
center (2tuple, optional) – Optional center of rotation. Origin is the upper left corner. Default is the center of the image.

torchvision.transforms.functional.
ten_crop
(img, size, vertical_flip=False)[source]¶  Crop the given PIL Image into four corners and the central crop plus the
flipped version of these (horizontal flipping is used by default).
Note
This transform returns a tuple of images and there may be a mismatch in the number of inputs and targets your
Dataset
returns. Parameters
 Returns
 tuple (tl, tr, bl, br, center, tl_flip, tr_flip, bl_flip, br_flip, center_flip)
Corresponding top left, top right, bottom left, bottom right and center crop and same for the flipped image.
 Return type

torchvision.transforms.functional.
to_grayscale
(img, num_output_channels=1)[source]¶ Convert image to grayscale version of image.
 Parameters
img (PIL Image) – Image to be converted to grayscale.
 Returns
 Grayscale version of the image.
if num_output_channels = 1 : returned image is single channel
if num_output_channels = 3 : returned image is 3 channel with r = g = b
 Return type
PIL Image

torchvision.transforms.functional.
to_pil_image
(pic, mode=None)[source]¶ Convert a tensor or an ndarray to PIL Image.
See
ToPILImage
for more details. Parameters
pic (Tensor or numpy.ndarray) – Image to be converted to PIL Image.
mode (PIL.Image mode) – color space and pixel depth of input data (optional).
 Returns
Image converted to PIL Image.
 Return type
PIL Image

torchvision.transforms.functional.
to_tensor
(pic)[source]¶ Convert a
PIL Image
ornumpy.ndarray
to tensor.See
ToTensor
for more details. Parameters
pic (PIL Image or numpy.ndarray) – Image to be converted to tensor.
 Returns
Converted image.
 Return type