Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[BC-breaking] Introduced InterpolationModes and deprecated arguments: resample and fillcolor #2952

Merged
merged 16 commits into from
Nov 27, 2020
Merged

[BC-breaking] Introduced InterpolationModes and deprecated arguments: resample and fillcolor #2952

Show file tree
Hide file tree
Changes from 11 commits
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
1b22ddc
Deprecated arguments: resample and fillcolor
vfdev-5 Nov 3, 2020
20c7177
Updates according to the review
vfdev-5 Nov 3, 2020
7f621a3
Merge branch 'master' into vfdev-5/deprecate-resample-fillcolor
vfdev-5 Nov 3, 2020
1bc499e
Added tests to check warnings and asserted BC
vfdev-5 Nov 3, 2020
f0d0257
Merge branch 'master' into vfdev-5/deprecate-resample-fillcolor
vfdev-5 Nov 5, 2020
453722c
Merge branch 'master' into vfdev-5/deprecate-resample-fillcolor
vfdev-5 Nov 5, 2020
d9c48d6
Merge branch 'master' of github.com:pytorch/vision into vfdev-5/depre…
vfdev-5 Nov 10, 2020
665350e
[WIP] Interpolation modes
vfdev-5 Nov 16, 2020
ceadb65
Added InterpolationModes enum
vfdev-5 Nov 16, 2020
e609722
Merge branch 'vfdev-5/deprecate-resample-fillcolor' of github.com:Qua…
vfdev-5 Nov 16, 2020
ed1b509
Merge branch 'master' into vfdev-5/deprecate-resample-fillcolor
vfdev-5 Nov 16, 2020
3768519
Added supported for int values for interpolation for BC
vfdev-5 Nov 20, 2020
6335042
Merge branch 'master' into vfdev-5/deprecate-resample-fillcolor
vfdev-5 Nov 20, 2020
4fb7790
Merge branch 'vfdev-5/deprecate-resample-fillcolor' of github.com:Qua…
vfdev-5 Nov 20, 2020
cf3d711
Removed useless test code
vfdev-5 Nov 20, 2020
6907505
Fix flake8
vfdev-5 Nov 20, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
73 changes: 48 additions & 25 deletions test/test_functional_tensor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,16 +4,19 @@
import math

import numpy as np
from PIL.Image import NEAREST, BILINEAR, BICUBIC

import torch
import torchvision.transforms.functional_tensor as F_t
import torchvision.transforms.functional_pil as F_pil
import torchvision.transforms.functional as F
from torchvision.transforms import InterpolationModes

from common_utils import TransformsTester


NEAREST, BILINEAR, BICUBIC = InterpolationModes.NEAREST, InterpolationModes.BILINEAR, InterpolationModes.BICUBIC


class Tester(TransformsTester):

def setUp(self):
Expand Down Expand Up @@ -365,7 +368,7 @@ def test_adjust_gamma(self):
)

def test_resize(self):
script_fn = torch.jit.script(F_t.resize)
script_fn = torch.jit.script(F.resize)
tensor, pil_img = self._create_data(26, 36, device=self.device)
batch_tensors = self._create_data_batch(16, 18, num_samples=4, device=self.device)

Expand All @@ -382,8 +385,8 @@ def test_resize(self):

for size in [32, 26, [32, ], [32, 32], (32, 32), [26, 35]]:
for interpolation in [BILINEAR, BICUBIC, NEAREST]:
resized_tensor = F_t.resize(tensor, size=size, interpolation=interpolation)
resized_pil_img = F_pil.resize(pil_img, size=size, interpolation=interpolation)
resized_tensor = F.resize(tensor, size=size, interpolation=interpolation)
resized_pil_img = F.resize(pil_img, size=size, interpolation=interpolation)

self.assertEqual(
resized_tensor.size()[1:], resized_pil_img.size[::-1], msg="{}, {}".format(size, interpolation)
Expand Down Expand Up @@ -418,13 +421,13 @@ def test_resized_crop(self):
# test values of F.resized_crop in several cases:
# 1) resize to the same size, crop to the same size => should be identity
tensor, _ = self._create_data(26, 36, device=self.device)
for i in [0, 2, 3]:
out_tensor = F.resized_crop(tensor, top=0, left=0, height=26, width=36, size=[26, 36], interpolation=i)
for mode in [NEAREST, BILINEAR, BICUBIC]:
out_tensor = F.resized_crop(tensor, top=0, left=0, height=26, width=36, size=[26, 36], interpolation=mode)
self.assertTrue(tensor.equal(out_tensor), msg="{} vs {}".format(out_tensor[0, :5, :5], tensor[0, :5, :5]))

# 2) resize by half and crop a TL corner
tensor, _ = self._create_data(26, 36, device=self.device)
out_tensor = F.resized_crop(tensor, top=0, left=0, height=20, width=30, size=[10, 15], interpolation=0)
out_tensor = F.resized_crop(tensor, top=0, left=0, height=20, width=30, size=[10, 15], interpolation=NEAREST)
expected_out_tensor = tensor[:, :20:2, :30:2]
self.assertTrue(
expected_out_tensor.equal(out_tensor),
Expand All @@ -433,17 +436,19 @@ def test_resized_crop(self):

batch_tensors = self._create_data_batch(26, 36, num_samples=4, device=self.device)
self._test_fn_on_batch(
batch_tensors, F.resized_crop, top=1, left=2, height=20, width=30, size=[10, 15], interpolation=0
batch_tensors, F.resized_crop, top=1, left=2, height=20, width=30, size=[10, 15], interpolation=NEAREST
)

def _test_affine_identity_map(self, tensor, scripted_affine):
# 1) identity map
out_tensor = F.affine(tensor, angle=0, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], resample=0)
out_tensor = F.affine(tensor, angle=0, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST)

self.assertTrue(
tensor.equal(out_tensor), msg="{} vs {}".format(out_tensor[0, :5, :5], tensor[0, :5, :5])
)
out_tensor = scripted_affine(tensor, angle=0, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], resample=0)
out_tensor = scripted_affine(
tensor, angle=0, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST
)
self.assertTrue(
tensor.equal(out_tensor), msg="{} vs {}".format(out_tensor[0, :5, :5], tensor[0, :5, :5])
)
Expand All @@ -461,13 +466,13 @@ def _test_affine_square_rotations(self, tensor, pil_img, scripted_affine):
]
for a, true_tensor in test_configs:
out_pil_img = F.affine(
pil_img, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], resample=0
pil_img, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST
)
out_pil_tensor = torch.from_numpy(np.array(out_pil_img).transpose((2, 0, 1))).to(self.device)

for fn in [F.affine, scripted_affine]:
out_tensor = fn(
tensor, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], resample=0
tensor, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST
)
if true_tensor is not None:
self.assertTrue(
Expand Down Expand Up @@ -496,13 +501,13 @@ def _test_affine_rect_rotations(self, tensor, pil_img, scripted_affine):
for a in test_configs:

out_pil_img = F.affine(
pil_img, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], resample=0
pil_img, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST
)
out_pil_tensor = torch.from_numpy(np.array(out_pil_img).transpose((2, 0, 1)))

for fn in [F.affine, scripted_affine]:
out_tensor = fn(
tensor, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], resample=0
tensor, angle=a, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST
).cpu()

if out_tensor.dtype != torch.uint8:
Expand All @@ -526,10 +531,10 @@ def _test_affine_translations(self, tensor, pil_img, scripted_affine):
]
for t in test_configs:

out_pil_img = F.affine(pil_img, angle=0, translate=t, scale=1.0, shear=[0.0, 0.0], resample=0)
out_pil_img = F.affine(pil_img, angle=0, translate=t, scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST)

for fn in [F.affine, scripted_affine]:
out_tensor = fn(tensor, angle=0, translate=t, scale=1.0, shear=[0.0, 0.0], resample=0)
out_tensor = fn(tensor, angle=0, translate=t, scale=1.0, shear=[0.0, 0.0], interpolation=NEAREST)

if out_tensor.dtype != torch.uint8:
out_tensor = out_tensor.to(torch.uint8)
Expand All @@ -550,13 +555,13 @@ def _test_affine_all_ops(self, tensor, pil_img, scripted_affine):
(-45, [-10, -10], 1.2, [4.0, 5.0]),
(-90, [0, 0], 1.0, [0.0, 0.0]),
]
for r in [0, ]:
for r in [NEAREST, ]:
for a, t, s, sh in test_configs:
out_pil_img = F.affine(pil_img, angle=a, translate=t, scale=s, shear=sh, resample=r)
out_pil_img = F.affine(pil_img, angle=a, translate=t, scale=s, shear=sh, interpolation=r)
out_pil_tensor = torch.from_numpy(np.array(out_pil_img).transpose((2, 0, 1)))

for fn in [F.affine, scripted_affine]:
out_tensor = fn(tensor, angle=a, translate=t, scale=s, shear=sh, resample=r).cpu()
out_tensor = fn(tensor, angle=a, translate=t, scale=s, shear=sh, interpolation=r).cpu()

if out_tensor.dtype != torch.uint8:
out_tensor = out_tensor.to(torch.uint8)
Expand Down Expand Up @@ -605,18 +610,30 @@ def test_affine(self):
batch_tensors, F.affine, angle=-43, translate=[-3, 4], scale=1.2, shear=[4.0, 5.0]
)

tensor, pil_img = data[0]
# assert deprecation warning and non-BC
with self.assertWarnsRegex(UserWarning, r"Argument resample is deprecated and will be removed"):
res1 = F.affine(tensor, 45, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], resample=2)
res2 = F.affine(tensor, 45, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], interpolation=BILINEAR)
self.assertTrue(res1.equal(res2))

with self.assertWarnsRegex(UserWarning, r"Argument fillcolor is deprecated and will be removed"):
res1 = F.affine(pil_img, 45, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], fillcolor=10)
res2 = F.affine(pil_img, 45, translate=[0, 0], scale=1.0, shear=[0.0, 0.0], fill=10)
self.assertEqual(res1, res2)

def _test_rotate_all_options(self, tensor, pil_img, scripted_rotate, centers):
img_size = pil_img.size
dt = tensor.dtype
for r in [0, ]:
for r in [NEAREST, ]:
for a in range(-180, 180, 17):
for e in [True, False]:
for c in centers:

out_pil_img = F.rotate(pil_img, angle=a, resample=r, expand=e, center=c)
out_pil_img = F.rotate(pil_img, angle=a, interpolation=r, expand=e, center=c)
out_pil_tensor = torch.from_numpy(np.array(out_pil_img).transpose((2, 0, 1)))
for fn in [F.rotate, scripted_rotate]:
out_tensor = fn(tensor, angle=a, resample=r, expand=e, center=c).cpu()
out_tensor = fn(tensor, angle=a, interpolation=r, expand=e, center=c).cpu()

if out_tensor.dtype != torch.uint8:
out_tensor = out_tensor.to(torch.uint8)
Expand Down Expand Up @@ -673,12 +690,18 @@ def test_rotate(self):

center = (20, 22)
self._test_fn_on_batch(
batch_tensors, F.rotate, angle=32, resample=0, expand=True, center=center
batch_tensors, F.rotate, angle=32, interpolation=NEAREST, expand=True, center=center
)
tensor, pil_img = data[0]
# assert deprecation warning and non-BC
with self.assertWarnsRegex(UserWarning, r"Argument resample is deprecated and will be removed"):
res1 = F.rotate(tensor, 45, resample=2)
res2 = F.rotate(tensor, 45, interpolation=BILINEAR)
self.assertTrue(res1.equal(res2))

def _test_perspective(self, tensor, pil_img, scripted_transform, test_configs):
dt = tensor.dtype
for r in [0, ]:
for r in [NEAREST, ]:
for spoints, epoints in test_configs:
out_pil_img = F.perspective(pil_img, startpoints=spoints, endpoints=epoints, interpolation=r)
out_pil_tensor = torch.from_numpy(np.array(out_pil_img).transpose((2, 0, 1)))
Expand Down Expand Up @@ -739,7 +762,7 @@ def test_perspective(self):

for spoints, epoints in test_configs:
self._test_fn_on_batch(
batch_tensors, F.perspective, startpoints=spoints, endpoints=epoints, interpolation=0
batch_tensors, F.perspective, startpoints=spoints, endpoints=epoints, interpolation=NEAREST
)

def test_gaussian_blur(self):
Expand Down
20 changes: 17 additions & 3 deletions test/test_transforms.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1484,11 +1484,16 @@ def test_random_rotation(self):

t = transforms.RandomRotation((-10, 10))
angle = t.get_params(t.degrees)
self.assertTrue(angle > -10 and angle < 10)
self.assertTrue(-10 < angle < 10)

# Checking if RandomRotation can be printed as string
t.__repr__()

# assert deprecation warning and non-BC
with self.assertWarnsRegex(UserWarning, r"Argument resample is deprecated and will be removed"):
t = transforms.RandomRotation((-10, 10), resample=2)
self.assertEqual(t.interpolation, transforms.InterpolationModes.BILINEAR)

def test_random_affine(self):

with self.assertRaises(ValueError):
Expand Down Expand Up @@ -1529,8 +1534,17 @@ def test_random_affine(self):
# Checking if RandomAffine can be printed as string
t.__repr__()

t = transforms.RandomAffine(10, resample=Image.BILINEAR)
self.assertIn("Image.BILINEAR", t.__repr__())
t = transforms.RandomAffine(10, interpolation=transforms.InterpolationModes.BILINEAR)
self.assertIn("bilinear", t.__repr__())

# assert deprecation warning and non-BC
with self.assertWarnsRegex(UserWarning, r"Argument resample is deprecated and will be removed"):
t = transforms.RandomAffine(10, resample=2)
self.assertEqual(t.interpolation, transforms.InterpolationModes.BILINEAR)

with self.assertWarnsRegex(UserWarning, r"Argument fillcolor is deprecated and will be removed"):
t = transforms.RandomAffine(10, fillcolor=10)
self.assertEqual(t.fill, 10)

def test_to_grayscale(self):
"""Unit tests for grayscale transform"""
Expand Down
10 changes: 6 additions & 4 deletions test/test_transforms_tensor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,7 @@
import torch
from torchvision import transforms as T
from torchvision.transforms import functional as F

from PIL.Image import NEAREST, BILINEAR, BICUBIC
from torchvision.transforms import InterpolationModes

import numpy as np

Expand All @@ -12,6 +11,9 @@
from common_utils import TransformsTester, get_tmp_dir, int_dtypes, float_dtypes


NEAREST, BILINEAR, BICUBIC = InterpolationModes.NEAREST, InterpolationModes.BILINEAR, InterpolationModes.BICUBIC


class Tester(TransformsTester):

def setUp(self):
Expand Down Expand Up @@ -349,7 +351,7 @@ def test_random_affine(self):
for interpolation in [NEAREST, BILINEAR]:
transform = T.RandomAffine(
degrees=degrees, translate=translate,
scale=scale, shear=shear, resample=interpolation
scale=scale, shear=shear, interpolation=interpolation
)
s_transform = torch.jit.script(transform)

Expand All @@ -368,7 +370,7 @@ def test_random_rotate(self):
for degrees in [45, 35.0, (-45, 45), [-90.0, 90.0]]:
for interpolation in [NEAREST, BILINEAR]:
transform = T.RandomRotation(
degrees=degrees, resample=interpolation, expand=expand, center=center
degrees=degrees, interpolation=interpolation, expand=expand, center=center
)
s_transform = torch.jit.script(transform)

Expand Down
Loading