GitLab

* Improved error messages for transforms.functional.normalize().

Split the original TypeError into 1. a TypeError if `tensor` is not a torch.Tensor and 2. a ValueError if `tensor` does not have the correct dimensionality.

Added more detail to the error for when `tensor` has the wrong dimension to make it easier to diagnose. This is useful when this function isn't called directly by the user (e.g. when the user uses transforms.Normalize and can't directly see this functions doc string).

Deleted hanging function `_is_tensor_image()`. It isn't used in this file and isn't used internally anywhere else in torchvision that I can see. (Some users will have used it despite the underscore prefix, but a quick google search for "F._is_tensor_image" suggests this is rare).

* Value checking to prevent division by zero runtime crashes.

Added a ValueError to check for and avoid division by zero in `div_`. Not preventing the call leads to runtime crashes, at least in some environments.

* Fixed div by zero check for non-scalar inputs.

I have fixed a typo that was persent in the Normalize class changed line 150 from     ``input[channel] = (input[channel] - mean[channel]) / std[channel]`` to     ``output[channel] = (input[channel] - mean[channel]) / std[channel]``

* initial fix

* outsourced num bands lookup

* fix doc

* added pillow version requirement

* simplify number of bands extraction

* remove unrelated change

* remove indirect dependency on pillow>=5.2.0

* extend docstring to transform

* bug fix

* added test

* scriptability checks

* tests upds

* linter upds

* linter upds

* upds

* tuple list changes

* linter updates

* add scriptable transform: center_crop

* add test: center_crop

* add scriptable transform: five_crop

* add scriptable transform: five_crop

* add scriptable transform: fix minor issues

* Add Scriptable Transform: Grayscale

* add scriptable transforms: rgb_to_grayscale

* add scriptable transform: rgb_to_grayscale

* add scriptable transform: rgb_to_grayscale

* add scriptable transform: rgb_to_grayscale

* update code: rgb_to_grayscale

* add test: rgb_to_grayscale

* update parameters: rgb_to_grayscale

* add scriptable transform: rgb_to_grayscale

* update rgb_to_grayscale

* update rgb_to_grayscale

* Improve readability of affine transformation code

* Make shear transformation area preserving

The previous shear implementation did not preserve area, and we
implement a version that does.

The formula used was verified with the following sympy code:

from sympy import Matrix, cos, sin, tan, simplify
from sympy.abc import x, y, phi

Xs = Matrix(
        [[1, -tan(x)],
         [0, 1]]
        )

Ys = Matrix(
        [[1, 0],
         [-tan(y), 1]]
        )

R = Matrix(
        [[cos(phi), -sin(phi)],
         [sin(phi), cos(phi)]]
        )

RSS = Matrix(
        [[cos(phi - y)/cos(y), -cos(phi - y)*tan(x)/cos(y) - sin(phi)],
         [sin(phi - y)/cos(y), -sin(phi - y)*tan(x)/cos(y) + cos(phi)]])

print(simplify(R * Ys * Xs - RSS))

One thing that is not clear (and could be tested) is whether avoiding
the explicit products and calculations in _get_inverse_affine_matrix
really gives performance benefits - compared to doing the explicit
calculation done in _test_transformation.

* Use np.matmul instead of @

The @ syntax is not supported in Python 2.

* Add adjustment operations for RGB Tensor Images.

Right now, we have operations on PIL images, but we want to have a version of the opeartions that act directly on Tensor images.

Here, we add such operations for adjust_brightness, adjust_contrast and adjust_saturation.

In PIL, those functions are implemented by generating an degenerate image from the first, and then interpolating them together.
- https://github.com/python-pillow/Pillow/blob/master/src/PIL/ImageEnhance.py
- https://github.com/python-pillow/Pillow/blob/master/src/libImaging/Blend.c

A few caveats:
* Since PIL operates on uint8, and the tensor operations might be on float, we can get slightly different values because of int truncation.
* We assume here the images are RGB; in particular, to handle an alpha channel, we need to check whether it is present, in which case we copy it to the final image.

* Keep dtype and use broadcast in adjust operations

- We make our operations have input.dtype == output.dtype, at the cost of
adding a few type checks and branches.

- By using Tensor broadcast, we can simplify the calls to _blend.

* Use is_floating_point to check dtype.

* Remove unpacking in tuple

It seems Python 2 does not support this type of unpacking, so it broke
Python 2 builds. This should fix it.

* Add from __future__ import division for Python 2

* Make crop torchscriptable

Relevant #1375

* Invert x and y axis

* fix lint

* Add crop test

* revert deletion of space in functional

* add import random

* add dimension in doc

* add import

* fix flake8

* change to self.assert*

* convert to uint8

* assertTrue

* lint

* doc-build fixed

* deprecation fixes

* deprecation updates

* vflip and hflip tensor

* vflip and hflip tensor

* changes made

* lint

* lint

* lint failing

* video transforms

* [video transforms]in ToTensorVideo, divide value by 255.0

* [video transforms] fix a bug

* fix linting

* Make changes backwards-compatible

* Adds optional fill colour to rotate

* bug fix

* Fix flakiness of test_randomresized_params

* Real fix

* Reduce number of iters

* to_pil_image updates

* lint

* Update test_transforms.py

* Update test_transforms.py

* Fix bug to Random Erasing

1. Avoid forever loop for getting parameters of erase.
2. replace' img_b' by 'img_c', because it indicates the channel.
3. replace v = torch.rand([img_c, h, w]) by v = torch.empty([img_c, h, w], dtype=torch.float32).normal_(). Normally distributed achieves better performance.

* add test

* Update test_transforms.py

* Update transforms.py

* Update test_transforms.py

* Update transforms.py

* Update functional.py

* Minor optimization to RandomErasing

This PR adds an additional check on `p` argument and prevents computing `img.shape` multiple times.

* linting

* initial commit

* add more checks, fix lint, fix doc

* updates on normalize

* test fixes

* Update test_transforms.py

* test improved

* Update test_transforms.py

* behaviour changes RandomErasing

* test fixes

* linter fix

* add erase function

* add Random Erasing

* Update transforms.py

* Update transforms.py

* add test for random erasing

* Update test_transforms.py

* fix flake8

* Update test_transforms.py

* Update functional.py

* Update test_transforms.py

* fix bug for per-pixel erasing

* Update transforms.py

* specific for coordinate (x, y)

* add raise TypeError for img

* Update transforms.py

* Update transforms.rst

* fix:error message of to_tensor

The error "pic should be PIL Image or ndarray. Got '<numpy.ndarray>'" is confusing.

* fix:a clearer function name

_is_numpy_image is clearer than _is_numpy_image_dim

* fix:add a test case

Add a test case in test/test_transforms.py to test the error message

* fix:pass ci check

* fix:wrong random matrix

* Fix normalize for different dtype than float32

* Fix lint

Change documentation of perspective(). The doc was about an old version that used directly transformation parameters

* updating docs for randomperspective

* my

* ci

Enable fillcolor option for affine transformation for Pillow >= 5.0.0 as described

* final changes

* final

* linter

* test changes

* linter

* lint

* indent

* lint

* minor changes

* parameter added

* ci

* ci fixes

* indent

* indent

* indent

* arg fixed

* Update test_transforms.py

* Update transforms.py

Changed 'channely' to 'channel' in the comments of normalize function.