test_transforms_v2.py 50.8 KB
Newer Older
1
2
3
import itertools
import pathlib
import random
4
import textwrap
5
6
7
8
9
10
11
12
13
import warnings

import numpy as np

import PIL.Image
import pytest
import torch
import torchvision.transforms.v2 as transforms

14
15
16
17
18
19
20
21
from common_utils import assert_equal, assert_run_python_script, cpu_and_cuda
from torch.utils._pytree import tree_flatten, tree_unflatten
from torchvision import datapoints
from torchvision.ops.boxes import box_iou
from torchvision.transforms.functional import to_pil_image
from torchvision.transforms.v2 import functional as F
from torchvision.transforms.v2.utils import check_type, is_pure_tensor, query_chw
from transforms_v2_legacy_utils import (
22
23
24
25
    make_bounding_boxes,
    make_detection_mask,
    make_image,
    make_images,
26
    make_multiple_bounding_boxes,
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
    make_segmentation_mask,
    make_video,
    make_videos,
)


def make_vanilla_tensor_images(*args, **kwargs):
    for image in make_images(*args, **kwargs):
        if image.ndim > 3:
            continue
        yield image.data


def make_pil_images(*args, **kwargs):
    for image in make_vanilla_tensor_images(*args, **kwargs):
        yield to_pil_image(image)


def make_vanilla_tensor_bounding_boxes(*args, **kwargs):
46
    for bounding_boxes in make_multiple_bounding_boxes(*args, **kwargs):
47
        yield bounding_boxes.data
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68


def parametrize(transforms_with_inputs):
    return pytest.mark.parametrize(
        ("transform", "input"),
        [
            pytest.param(
                transform,
                input,
                id=f"{type(transform).__name__}-{type(input).__module__}.{type(input).__name__}-{idx}",
            )
            for transform, inputs in transforms_with_inputs
            for idx, input in enumerate(inputs)
        ],
    )


def auto_augment_adapter(transform, input, device):
    adapted_input = {}
    image_or_video_found = False
    for key, value in input.items():
69
        if isinstance(value, (datapoints.BoundingBoxes, datapoints.Mask)):
70
71
            # AA transforms don't support bounding boxes or masks
            continue
72
        elif check_type(value, (datapoints.Image, datapoints.Video, is_pure_tensor, PIL.Image.Image)):
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
            if image_or_video_found:
                # AA transforms only support a single image or video
                continue
            image_or_video_found = True
        adapted_input[key] = value
    return adapted_input


def linear_transformation_adapter(transform, input, device):
    flat_inputs = list(input.values())
    c, h, w = query_chw(
        [
            item
            for item, needs_transform in zip(flat_inputs, transforms.Transform()._needs_transform_list(flat_inputs))
            if needs_transform
        ]
    )
    num_elements = c * h * w
    transform.transformation_matrix = torch.randn((num_elements, num_elements), device=device)
    transform.mean_vector = torch.randn((num_elements,), device=device)
    return {key: value for key, value in input.items() if not isinstance(value, PIL.Image.Image)}


def normalize_adapter(transform, input, device):
    adapted_input = {}
    for key, value in input.items():
        if isinstance(value, PIL.Image.Image):
            # normalize doesn't support PIL images
            continue
102
        elif check_type(value, (datapoints.Image, datapoints.Video, is_pure_tensor)):
103
            # normalize doesn't support integer images
104
            value = F.to_dtype(value, torch.float32, scale=True)
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
        adapted_input[key] = value
    return adapted_input


class TestSmoke:
    @pytest.mark.parametrize(
        ("transform", "adapter"),
        [
            (transforms.RandomErasing(p=1.0), None),
            (transforms.AugMix(), auto_augment_adapter),
            (transforms.AutoAugment(), auto_augment_adapter),
            (transforms.RandAugment(), auto_augment_adapter),
            (transforms.TrivialAugmentWide(), auto_augment_adapter),
            (transforms.ColorJitter(brightness=0.1, contrast=0.2, saturation=0.3, hue=0.15), None),
            (transforms.Grayscale(), None),
            (transforms.RandomAdjustSharpness(sharpness_factor=0.5, p=1.0), None),
            (transforms.RandomAutocontrast(p=1.0), None),
            (transforms.RandomEqualize(p=1.0), None),
            (transforms.RandomGrayscale(p=1.0), None),
            (transforms.RandomInvert(p=1.0), None),
125
            (transforms.RandomChannelPermutation(), None),
126
127
128
129
130
131
132
133
134
135
            (transforms.RandomPhotometricDistort(p=1.0), None),
            (transforms.RandomPosterize(bits=4, p=1.0), None),
            (transforms.RandomSolarize(threshold=0.5, p=1.0), None),
            (transforms.CenterCrop([16, 16]), None),
            (transforms.ElasticTransform(sigma=1.0), None),
            (transforms.Pad(4), None),
            (transforms.RandomAffine(degrees=30.0), None),
            (transforms.RandomCrop([16, 16], pad_if_needed=True), None),
            (transforms.RandomHorizontalFlip(p=1.0), None),
            (transforms.RandomPerspective(p=1.0), None),
136
137
            (transforms.RandomResize(min_size=10, max_size=20, antialias=True), None),
            (transforms.RandomResizedCrop([16, 16], antialias=True), None),
138
            (transforms.RandomRotation(degrees=30), None),
139
            (transforms.RandomShortestSize(min_size=10, antialias=True), None),
140
141
142
            (transforms.RandomVerticalFlip(p=1.0), None),
            (transforms.RandomZoomOut(p=1.0), None),
            (transforms.Resize([16, 16], antialias=True), None),
143
            (transforms.ScaleJitter((16, 16), scale_range=(0.8, 1.2), antialias=True), None),
144
            (transforms.ClampBoundingBoxes(), None),
145
            (transforms.ConvertBoundingBoxFormat(datapoints.BoundingBoxFormat.CXCYWH), None),
146
            (transforms.ConvertImageDtype(), None),
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
            (transforms.GaussianBlur(kernel_size=3), None),
            (
                transforms.LinearTransformation(
                    # These are just dummy values that will be filled by the adapter. We can't define them upfront,
                    # because for we neither know the spatial size nor the device at this point
                    transformation_matrix=torch.empty((1, 1)),
                    mean_vector=torch.empty((1,)),
                ),
                linear_transformation_adapter,
            ),
            (transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]), normalize_adapter),
            (transforms.ToDtype(torch.float64), None),
            (transforms.UniformTemporalSubsample(num_samples=2), None),
        ],
        ids=lambda transform: type(transform).__name__,
    )
    @pytest.mark.parametrize("container_type", [dict, list, tuple])
    @pytest.mark.parametrize(
        "image_or_video",
        [
            make_image(),
            make_video(),
            next(make_pil_images(color_spaces=["RGB"])),
            next(make_vanilla_tensor_images()),
        ],
    )
173
    @pytest.mark.parametrize("device", cpu_and_cuda())
174
    def test_common(self, transform, adapter, container_type, image_or_video, device):
Philip Meier's avatar
Philip Meier committed
175
        canvas_size = F.get_size(image_or_video)
176
177
        input = dict(
            image_or_video=image_or_video,
Philip Meier's avatar
Philip Meier committed
178
179
180
            image_datapoint=make_image(size=canvas_size),
            video_datapoint=make_video(size=canvas_size),
            image_pil=next(make_pil_images(sizes=[canvas_size], color_spaces=["RGB"])),
181
            bounding_boxes_xyxy=make_bounding_boxes(
Philip Meier's avatar
Philip Meier committed
182
                format=datapoints.BoundingBoxFormat.XYXY, canvas_size=canvas_size, batch_dims=(3,)
183
            ),
184
            bounding_boxes_xywh=make_bounding_boxes(
Philip Meier's avatar
Philip Meier committed
185
                format=datapoints.BoundingBoxFormat.XYWH, canvas_size=canvas_size, batch_dims=(4,)
186
            ),
187
            bounding_boxes_cxcywh=make_bounding_boxes(
Philip Meier's avatar
Philip Meier committed
188
                format=datapoints.BoundingBoxFormat.CXCYWH, canvas_size=canvas_size, batch_dims=(5,)
189
            ),
190
            bounding_boxes_degenerate_xyxy=datapoints.BoundingBoxes(
191
192
193
194
195
196
197
198
199
                [
                    [0, 0, 0, 0],  # no height or width
                    [0, 0, 0, 1],  # no height
                    [0, 0, 1, 0],  # no width
                    [2, 0, 1, 1],  # x1 > x2, y1 < y2
                    [0, 2, 1, 1],  # x1 < x2, y1 > y2
                    [2, 2, 1, 1],  # x1 > x2, y1 > y2
                ],
                format=datapoints.BoundingBoxFormat.XYXY,
Philip Meier's avatar
Philip Meier committed
200
                canvas_size=canvas_size,
201
            ),
202
            bounding_boxes_degenerate_xywh=datapoints.BoundingBoxes(
203
204
205
206
207
208
209
210
211
                [
                    [0, 0, 0, 0],  # no height or width
                    [0, 0, 0, 1],  # no height
                    [0, 0, 1, 0],  # no width
                    [0, 0, 1, -1],  # negative height
                    [0, 0, -1, 1],  # negative width
                    [0, 0, -1, -1],  # negative height and width
                ],
                format=datapoints.BoundingBoxFormat.XYWH,
Philip Meier's avatar
Philip Meier committed
212
                canvas_size=canvas_size,
213
            ),
214
            bounding_boxes_degenerate_cxcywh=datapoints.BoundingBoxes(
215
216
217
218
219
220
221
222
223
                [
                    [0, 0, 0, 0],  # no height or width
                    [0, 0, 0, 1],  # no height
                    [0, 0, 1, 0],  # no width
                    [0, 0, 1, -1],  # negative height
                    [0, 0, -1, 1],  # negative width
                    [0, 0, -1, -1],  # negative height and width
                ],
                format=datapoints.BoundingBoxFormat.CXCYWH,
Philip Meier's avatar
Philip Meier committed
224
                canvas_size=canvas_size,
225
            ),
Philip Meier's avatar
Philip Meier committed
226
227
            detection_mask=make_detection_mask(size=canvas_size),
            segmentation_mask=make_segmentation_mask(size=canvas_size),
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
            int=0,
            float=0.0,
            bool=True,
            none=None,
            str="str",
            path=pathlib.Path.cwd(),
            object=object(),
            tensor=torch.empty(5),
            array=np.empty(5),
        )
        if adapter is not None:
            input = adapter(transform, input, device)

        if container_type in {tuple, list}:
            input = container_type(input.values())

        input_flat, input_spec = tree_flatten(input)
        input_flat = [item.to(device) if isinstance(item, torch.Tensor) else item for item in input_flat]
        input = tree_unflatten(input_flat, input_spec)

        torch.manual_seed(0)
        output = transform(input)
        output_flat, output_spec = tree_flatten(output)

        assert output_spec == input_spec

        for output_item, input_item, should_be_transformed in zip(
            output_flat, input_flat, transforms.Transform()._needs_transform_list(input_flat)
        ):
            if should_be_transformed:
                assert type(output_item) is type(input_item)
            else:
                assert output_item is input_item

262
            if isinstance(input_item, datapoints.BoundingBoxes) and not isinstance(
263
264
265
266
267
268
269
270
271
                transform, transforms.ConvertBoundingBoxFormat
            ):
                assert output_item.format == input_item.format

        # Enforce that the transform does not turn a degenerate box marked by RandomIoUCrop (or any other future
        # transform that does this), back into a valid one.
        # TODO: we should test that against all degenerate boxes above
        for format in list(datapoints.BoundingBoxFormat):
            sample = dict(
Philip Meier's avatar
Philip Meier committed
272
                boxes=datapoints.BoundingBoxes([[0, 0, 0, 0]], format=format, canvas_size=(224, 244)),
273
274
                labels=torch.tensor([3]),
            )
275
            assert transforms.SanitizeBoundingBoxes()(sample)["boxes"].shape == (0, 4)
276
277
278
279
280
281
282
283
284
285
286
287
288

    @parametrize(
        [
            (
                transform,
                itertools.chain.from_iterable(
                    fn(
                        color_spaces=[
                            "GRAY",
                            "RGB",
                        ],
                        dtypes=[torch.uint8],
                        extra_dims=[(), (4,)],
289
                        **(dict(num_frames=[3]) if fn is make_videos else dict()),
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
                    )
                    for fn in [
                        make_images,
                        make_vanilla_tensor_images,
                        make_pil_images,
                        make_videos,
                    ]
                ),
            )
            for transform in (
                transforms.RandAugment(),
                transforms.TrivialAugmentWide(),
                transforms.AutoAugment(),
                transforms.AugMix(),
            )
        ]
    )
    def test_auto_augment(self, transform, input):
        transform(input)

    @parametrize(
        [
            (
                transforms.Normalize(mean=[0.0, 0.0, 0.0], std=[1.0, 1.0, 1.0]),
                itertools.chain.from_iterable(
                    fn(color_spaces=["RGB"], dtypes=[torch.float32])
                    for fn in [
                        make_images,
                        make_vanilla_tensor_images,
                        make_videos,
                    ]
                ),
            ),
        ]
    )
    def test_normalize(self, transform, input):
        transform(input)

    @parametrize(
        [
            (
                transforms.RandomResizedCrop([16, 16], antialias=True),
                itertools.chain(
                    make_images(extra_dims=[(4,)]),
                    make_vanilla_tensor_images(),
                    make_pil_images(),
                    make_videos(extra_dims=[()]),
                ),
            )
        ]
    )
    def test_random_resized_crop(self, transform, input):
        transform(input)


@pytest.mark.parametrize(
    "flat_inputs",
    itertools.permutations(
        [
            next(make_vanilla_tensor_images()),
            next(make_vanilla_tensor_images()),
            next(make_pil_images()),
            make_image(),
            next(make_videos()),
        ],
        3,
    ),
)
358
359
def test_pure_tensor_heuristic(flat_inputs):
    def split_on_pure_tensor(to_split):
360
        # This takes a sequence that is structurally aligned with `flat_inputs` and splits its items into three parts:
361
362
        # 1. The first pure tensor. If none is present, this will be `None`
        # 2. A list of the remaining pure tensors
363
        # 3. A list of all other items
364
        pure_tensors = []
365
366
367
368
        others = []
        # Splitting always happens on the original `flat_inputs` to avoid any erroneous type changes by the transform to
        # affect the splitting.
        for item, inpt in zip(to_split, flat_inputs):
369
370
            (pure_tensors if is_pure_tensor(inpt) else others).append(item)
        return pure_tensors[0] if pure_tensors else None, pure_tensors[1:], others
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385

    class CopyCloneTransform(transforms.Transform):
        def _transform(self, inpt, params):
            return inpt.clone() if isinstance(inpt, torch.Tensor) else inpt.copy()

        @staticmethod
        def was_applied(output, inpt):
            identity = output is inpt
            if identity:
                return False

            # Make sure nothing fishy is going on
            assert_equal(output, inpt)
            return True

386
    first_pure_tensor_input, other_pure_tensor_inputs, other_inputs = split_on_pure_tensor(flat_inputs)
387
388
389
390

    transform = CopyCloneTransform()
    transformed_sample = transform(flat_inputs)

391
    first_pure_tensor_output, other_pure_tensor_outputs, other_outputs = split_on_pure_tensor(transformed_sample)
392

393
    if first_pure_tensor_input is not None:
394
        if other_inputs:
395
            assert not transform.was_applied(first_pure_tensor_output, first_pure_tensor_input)
396
        else:
397
            assert transform.was_applied(first_pure_tensor_output, first_pure_tensor_input)
398

399
    for output, inpt in zip(other_pure_tensor_outputs, other_pure_tensor_inputs):
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
        assert not transform.was_applied(output, inpt)

    for input, output in zip(other_inputs, other_outputs):
        assert transform.was_applied(output, input)


class TestPad:
    def test_assertions(self):
        with pytest.raises(TypeError, match="Got inappropriate padding arg"):
            transforms.Pad("abc")

        with pytest.raises(ValueError, match="Padding must be an int or a 1, 2, or 4"):
            transforms.Pad([-0.7, 0, 0.7])

        with pytest.raises(TypeError, match="Got inappropriate fill arg"):
            transforms.Pad(12, fill="abc")

        with pytest.raises(ValueError, match="Padding mode should be either"):
            transforms.Pad(12, padding_mode="abc")


class TestRandomZoomOut:
    def test_assertions(self):
        with pytest.raises(TypeError, match="Got inappropriate fill arg"):
            transforms.RandomZoomOut(fill="abc")

        with pytest.raises(TypeError, match="should be a sequence of length"):
            transforms.RandomZoomOut(0, side_range=0)

        with pytest.raises(ValueError, match="Invalid canvas side range"):
            transforms.RandomZoomOut(0, side_range=[4.0, 1.0])

    @pytest.mark.parametrize("fill", [0, [1, 2, 3], (2, 3, 4)])
    @pytest.mark.parametrize("side_range", [(1.0, 4.0), [2.0, 5.0]])
Philip Meier's avatar
Philip Meier committed
434
    def test__get_params(self, fill, side_range):
435
436
        transform = transforms.RandomZoomOut(fill=fill, side_range=side_range)

Philip Meier's avatar
Philip Meier committed
437
438
        h, w = size = (24, 32)
        image = make_image(size)
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467

        params = transform._get_params([image])

        assert len(params["padding"]) == 4
        assert 0 <= params["padding"][0] <= (side_range[1] - 1) * w
        assert 0 <= params["padding"][1] <= (side_range[1] - 1) * h
        assert 0 <= params["padding"][2] <= (side_range[1] - 1) * w
        assert 0 <= params["padding"][3] <= (side_range[1] - 1) * h


class TestRandomCrop:
    def test_assertions(self):
        with pytest.raises(ValueError, match="Please provide only two dimensions"):
            transforms.RandomCrop([10, 12, 14])

        with pytest.raises(TypeError, match="Got inappropriate padding arg"):
            transforms.RandomCrop([10, 12], padding="abc")

        with pytest.raises(ValueError, match="Padding must be an int or a 1, 2, or 4"):
            transforms.RandomCrop([10, 12], padding=[-0.7, 0, 0.7])

        with pytest.raises(TypeError, match="Got inappropriate fill arg"):
            transforms.RandomCrop([10, 12], padding=1, fill="abc")

        with pytest.raises(ValueError, match="Padding mode should be either"):
            transforms.RandomCrop([10, 12], padding=1, padding_mode="abc")

    @pytest.mark.parametrize("padding", [None, 1, [2, 3], [1, 2, 3, 4]])
    @pytest.mark.parametrize("size, pad_if_needed", [((10, 10), False), ((50, 25), True)])
Philip Meier's avatar
Philip Meier committed
468
469
470
    def test__get_params(self, padding, pad_if_needed, size):
        h, w = size = (24, 32)
        image = make_image(size)
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549

        transform = transforms.RandomCrop(size, padding=padding, pad_if_needed=pad_if_needed)
        params = transform._get_params([image])

        if padding is not None:
            if isinstance(padding, int):
                pad_top = pad_bottom = pad_left = pad_right = padding
            elif isinstance(padding, list) and len(padding) == 2:
                pad_left = pad_right = padding[0]
                pad_top = pad_bottom = padding[1]
            elif isinstance(padding, list) and len(padding) == 4:
                pad_left, pad_top, pad_right, pad_bottom = padding

            h += pad_top + pad_bottom
            w += pad_left + pad_right
        else:
            pad_left = pad_right = pad_top = pad_bottom = 0

        if pad_if_needed:
            if w < size[1]:
                diff = size[1] - w
                pad_left += diff
                pad_right += diff
                w += 2 * diff
            if h < size[0]:
                diff = size[0] - h
                pad_top += diff
                pad_bottom += diff
                h += 2 * diff

        padding = [pad_left, pad_top, pad_right, pad_bottom]

        assert 0 <= params["top"] <= h - size[0] + 1
        assert 0 <= params["left"] <= w - size[1] + 1
        assert params["height"] == size[0]
        assert params["width"] == size[1]
        assert params["needs_pad"] is any(padding)
        assert params["padding"] == padding


class TestGaussianBlur:
    def test_assertions(self):
        with pytest.raises(ValueError, match="Kernel size should be a tuple/list of two integers"):
            transforms.GaussianBlur([10, 12, 14])

        with pytest.raises(ValueError, match="Kernel size value should be an odd and positive number"):
            transforms.GaussianBlur(4)

        with pytest.raises(
            TypeError, match="sigma should be a single int or float or a list/tuple with length 2 floats."
        ):
            transforms.GaussianBlur(3, sigma=[1, 2, 3])

        with pytest.raises(ValueError, match="If sigma is a single number, it must be positive"):
            transforms.GaussianBlur(3, sigma=-1.0)

        with pytest.raises(ValueError, match="sigma values should be positive and of the form"):
            transforms.GaussianBlur(3, sigma=[2.0, 1.0])

    @pytest.mark.parametrize("sigma", [10.0, [10.0, 12.0]])
    def test__get_params(self, sigma):
        transform = transforms.GaussianBlur(3, sigma=sigma)
        params = transform._get_params([])

        if isinstance(sigma, float):
            assert params["sigma"][0] == params["sigma"][1] == 10
        else:
            assert sigma[0] <= params["sigma"][0] <= sigma[1]
            assert sigma[0] <= params["sigma"][1] <= sigma[1]


class TestRandomPerspective:
    def test_assertions(self):
        with pytest.raises(ValueError, match="Argument distortion_scale value should be between 0 and 1"):
            transforms.RandomPerspective(distortion_scale=-1.0)

        with pytest.raises(TypeError, match="Got inappropriate fill arg"):
            transforms.RandomPerspective(0.5, fill="abc")

Philip Meier's avatar
Philip Meier committed
550
    def test__get_params(self):
551
552
        dscale = 0.5
        transform = transforms.RandomPerspective(dscale)
Philip Meier's avatar
Philip Meier committed
553
554

        image = make_image((24, 32))
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585

        params = transform._get_params([image])

        assert "coefficients" in params
        assert len(params["coefficients"]) == 8


class TestElasticTransform:
    def test_assertions(self):

        with pytest.raises(TypeError, match="alpha should be float or a sequence of floats"):
            transforms.ElasticTransform({})

        with pytest.raises(ValueError, match="alpha is a sequence its length should be one of 2"):
            transforms.ElasticTransform([1.0, 2.0, 3.0])

        with pytest.raises(ValueError, match="alpha should be a sequence of floats"):
            transforms.ElasticTransform([1, 2])

        with pytest.raises(TypeError, match="sigma should be float or a sequence of floats"):
            transforms.ElasticTransform(1.0, {})

        with pytest.raises(ValueError, match="sigma is a sequence its length should be one of 2"):
            transforms.ElasticTransform(1.0, [1.0, 2.0, 3.0])

        with pytest.raises(ValueError, match="sigma should be a sequence of floats"):
            transforms.ElasticTransform(1.0, [1, 2])

        with pytest.raises(TypeError, match="Got inappropriate fill arg"):
            transforms.ElasticTransform(1.0, 2.0, fill="abc")

Philip Meier's avatar
Philip Meier committed
586
    def test__get_params(self):
587
588
589
        alpha = 2.0
        sigma = 3.0
        transform = transforms.ElasticTransform(alpha, sigma)
Philip Meier's avatar
Philip Meier committed
590
591
592

        h, w = size = (24, 32)
        image = make_image(size)
593
594
595
596
597
598
599
600
601
602

        params = transform._get_params([image])

        displacement = params["displacement"]
        assert displacement.shape == (1, h, w, 2)
        assert (-alpha / w <= displacement[0, ..., 0]).all() and (displacement[0, ..., 0] <= alpha / w).all()
        assert (-alpha / h <= displacement[0, ..., 1]).all() and (displacement[0, ..., 1] <= alpha / h).all()


class TestRandomErasing:
Philip Meier's avatar
Philip Meier committed
603
    def test_assertions(self):
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
        with pytest.raises(TypeError, match="Argument value should be either a number or str or a sequence"):
            transforms.RandomErasing(value={})

        with pytest.raises(ValueError, match="If value is str, it should be 'random'"):
            transforms.RandomErasing(value="abc")

        with pytest.raises(TypeError, match="Scale should be a sequence"):
            transforms.RandomErasing(scale=123)

        with pytest.raises(TypeError, match="Ratio should be a sequence"):
            transforms.RandomErasing(ratio=123)

        with pytest.raises(ValueError, match="Scale should be between 0 and 1"):
            transforms.RandomErasing(scale=[-1, 2])

Philip Meier's avatar
Philip Meier committed
619
        image = make_image((24, 32))
620
621
622
623
624
625
626

        transform = transforms.RandomErasing(value=[1, 2, 3, 4])

        with pytest.raises(ValueError, match="If value is a sequence, it should have either a single value"):
            transform._get_params([image])

    @pytest.mark.parametrize("value", [5.0, [1, 2, 3], "random"])
Philip Meier's avatar
Philip Meier committed
627
628
629
    def test__get_params(self, value):
        image = make_image((24, 32))
        num_channels, height, width = F.get_dimensions(image)
630
631
632
633
634
635
636
637
638

        transform = transforms.RandomErasing(value=value)
        params = transform._get_params([image])

        v = params["v"]
        h, w = params["h"], params["w"]
        i, j = params["i"], params["j"]
        assert isinstance(v, torch.Tensor)
        if value == "random":
Philip Meier's avatar
Philip Meier committed
639
            assert v.shape == (num_channels, h, w)
640
641
642
        elif isinstance(value, (int, float)):
            assert v.shape == (1, 1, 1)
        elif isinstance(value, (list, tuple)):
Philip Meier's avatar
Philip Meier committed
643
            assert v.shape == (num_channels, 1, 1)
644

Philip Meier's avatar
Philip Meier committed
645
646
        assert 0 <= i <= height - h
        assert 0 <= j <= width - w
647
648
649
650
651


class TestTransform:
    @pytest.mark.parametrize(
        "inpt_type",
652
        [torch.Tensor, PIL.Image.Image, datapoints.Image, np.ndarray, datapoints.BoundingBoxes, str, int],
653
654
655
656
657
658
659
660
661
662
663
664
665
666
    )
    def test_check_transformed_types(self, inpt_type, mocker):
        # This test ensures that we correctly handle which types to transform and which to bypass
        t = transforms.Transform()
        inpt = mocker.MagicMock(spec=inpt_type)

        if inpt_type in (np.ndarray, str, int):
            output = t(inpt)
            assert output is inpt
        else:
            with pytest.raises(NotImplementedError):
                t(inpt)


667
class TestToImage:
668
669
    @pytest.mark.parametrize(
        "inpt_type",
670
        [torch.Tensor, PIL.Image.Image, datapoints.Image, np.ndarray, datapoints.BoundingBoxes, str, int],
671
672
673
    )
    def test__transform(self, inpt_type, mocker):
        fn = mocker.patch(
674
            "torchvision.transforms.v2.functional.to_image",
675
676
677
678
            return_value=torch.rand(1, 3, 8, 8),
        )

        inpt = mocker.MagicMock(spec=inpt_type)
679
        transform = transforms.ToImage()
680
        transform(inpt)
681
        if inpt_type in (datapoints.BoundingBoxes, datapoints.Image, str, int):
682
683
684
685
686
687
688
689
            assert fn.call_count == 0
        else:
            fn.assert_called_once_with(inpt)


class TestToPILImage:
    @pytest.mark.parametrize(
        "inpt_type",
690
        [torch.Tensor, PIL.Image.Image, datapoints.Image, np.ndarray, datapoints.BoundingBoxes, str, int],
691
692
    )
    def test__transform(self, inpt_type, mocker):
693
        fn = mocker.patch("torchvision.transforms.v2.functional.to_pil_image")
694
695
696
697

        inpt = mocker.MagicMock(spec=inpt_type)
        transform = transforms.ToPILImage()
        transform(inpt)
698
        if inpt_type in (PIL.Image.Image, datapoints.BoundingBoxes, str, int):
699
700
701
702
703
704
705
706
            assert fn.call_count == 0
        else:
            fn.assert_called_once_with(inpt, mode=transform.mode)


class TestToTensor:
    @pytest.mark.parametrize(
        "inpt_type",
707
        [torch.Tensor, PIL.Image.Image, datapoints.Image, np.ndarray, datapoints.BoundingBoxes, str, int],
708
709
710
711
712
713
714
715
    )
    def test__transform(self, inpt_type, mocker):
        fn = mocker.patch("torchvision.transforms.functional.to_tensor")

        inpt = mocker.MagicMock(spec=inpt_type)
        with pytest.warns(UserWarning, match="deprecated and will be removed"):
            transform = transforms.ToTensor()
        transform(inpt)
716
        if inpt_type in (datapoints.Image, torch.Tensor, datapoints.BoundingBoxes, str, int):
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
            assert fn.call_count == 0
        else:
            fn.assert_called_once_with(inpt)


class TestContainers:
    @pytest.mark.parametrize("transform_cls", [transforms.Compose, transforms.RandomChoice, transforms.RandomOrder])
    def test_assertions(self, transform_cls):
        with pytest.raises(TypeError, match="Argument transforms should be a sequence of callables"):
            transform_cls(transforms.RandomCrop(28))

    @pytest.mark.parametrize("transform_cls", [transforms.Compose, transforms.RandomChoice, transforms.RandomOrder])
    @pytest.mark.parametrize(
        "trfms",
        [
            [transforms.Pad(2), transforms.RandomCrop(28)],
            [lambda x: 2.0 * x, transforms.Pad(2), transforms.RandomCrop(28)],
            [transforms.Pad(2), lambda x: 2.0 * x, transforms.RandomCrop(28)],
        ],
    )
    def test_ctor(self, transform_cls, trfms):
        c = transform_cls(trfms)
        inpt = torch.rand(1, 3, 32, 32)
        output = c(inpt)
        assert isinstance(output, torch.Tensor)
        assert output.ndim == 4


class TestRandomChoice:
    def test_assertions(self):
747
        with pytest.raises(ValueError, match="Length of p doesn't match the number of transforms"):
748
            transforms.RandomChoice([transforms.Pad(2), transforms.RandomCrop(28)], p=[1])
749
750
751


class TestRandomIoUCrop:
752
    @pytest.mark.parametrize("device", cpu_and_cuda())
753
    @pytest.mark.parametrize("options", [[0.5, 0.9], [2.0]])
Philip Meier's avatar
Philip Meier committed
754
755
756
    def test__get_params(self, device, options):
        orig_h, orig_w = size = (24, 32)
        image = make_image(size)
757
        bboxes = datapoints.BoundingBoxes(
758
759
            torch.tensor([[1, 1, 10, 10], [20, 20, 23, 23], [1, 20, 10, 23], [20, 1, 23, 10]]),
            format="XYXY",
Philip Meier's avatar
Philip Meier committed
760
            canvas_size=size,
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
            device=device,
        )
        sample = [image, bboxes]

        transform = transforms.RandomIoUCrop(sampler_options=options)

        n_samples = 5
        for _ in range(n_samples):

            params = transform._get_params(sample)

            if options == [2.0]:
                assert len(params) == 0
                return

            assert len(params["is_within_crop_area"]) > 0
            assert params["is_within_crop_area"].dtype == torch.bool

            assert int(transform.min_scale * orig_h) <= params["height"] <= int(transform.max_scale * orig_h)
            assert int(transform.min_scale * orig_w) <= params["width"] <= int(transform.max_scale * orig_w)

            left, top = params["left"], params["top"]
            new_h, new_w = params["height"], params["width"]
            ious = box_iou(
                bboxes,
                torch.tensor([[left, top, left + new_w, top + new_h]], dtype=bboxes.dtype, device=bboxes.device),
            )
            assert ious.max() >= options[0] or ious.max() >= options[1], f"{ious} vs {options}"

    def test__transform_empty_params(self, mocker):
        transform = transforms.RandomIoUCrop(sampler_options=[2.0])
        image = datapoints.Image(torch.rand(1, 3, 4, 4))
Philip Meier's avatar
Philip Meier committed
793
        bboxes = datapoints.BoundingBoxes(torch.tensor([[1, 1, 2, 2]]), format="XYXY", canvas_size=(4, 4))
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
        label = torch.tensor([1])
        sample = [image, bboxes, label]
        # Let's mock transform._get_params to control the output:
        transform._get_params = mocker.MagicMock(return_value={})
        output = transform(sample)
        torch.testing.assert_close(output, sample)

    def test_forward_assertion(self):
        transform = transforms.RandomIoUCrop()
        with pytest.raises(
            TypeError,
            match="requires input sample to contain tensor or PIL images and bounding boxes",
        ):
            transform(torch.tensor(0))

    def test__transform(self, mocker):
        transform = transforms.RandomIoUCrop()

Philip Meier's avatar
Philip Meier committed
812
813
        size = (32, 24)
        image = make_image(size)
814
        bboxes = make_bounding_boxes(format="XYXY", canvas_size=size, batch_dims=(6,))
Philip Meier's avatar
Philip Meier committed
815
        masks = make_detection_mask(size, num_objects=6)
816
817
818
819
820
821
822
823
824
825
826

        sample = [image, bboxes, masks]

        is_within_crop_area = torch.tensor([0, 1, 0, 1, 0, 1], dtype=torch.bool)

        params = dict(top=1, left=2, height=12, width=12, is_within_crop_area=is_within_crop_area)
        transform._get_params = mocker.MagicMock(return_value=params)
        output = transform(sample)

        # check number of bboxes vs number of labels:
        output_bboxes = output[1]
827
        assert isinstance(output_bboxes, datapoints.BoundingBoxes)
828
829
830
831
832
833
834
        assert (output_bboxes[~is_within_crop_area] == 0).all()

        output_masks = output[2]
        assert isinstance(output_masks, datapoints.Mask)


class TestScaleJitter:
Philip Meier's avatar
Philip Meier committed
835
836
    def test__get_params(self):
        canvas_size = (24, 32)
837
838
839
840
        target_size = (16, 12)
        scale_range = (0.5, 1.5)

        transform = transforms.ScaleJitter(target_size=target_size, scale_range=scale_range)
Philip Meier's avatar
Philip Meier committed
841
842

        sample = make_image(canvas_size)
843
844
845
846
847
848
849
850
851
852
853
854

        n_samples = 5
        for _ in range(n_samples):

            params = transform._get_params([sample])

            assert "size" in params
            size = params["size"]

            assert isinstance(size, tuple) and len(size) == 2
            height, width = size

Philip Meier's avatar
Philip Meier committed
855
856
            r_min = min(target_size[1] / canvas_size[0], target_size[0] / canvas_size[1]) * scale_range[0]
            r_max = min(target_size[1] / canvas_size[0], target_size[0] / canvas_size[1]) * scale_range[1]
857

Philip Meier's avatar
Philip Meier committed
858
859
            assert int(canvas_size[0] * r_min) <= height <= int(canvas_size[0] * r_max)
            assert int(canvas_size[1] * r_min) <= width <= int(canvas_size[1] * r_max)
860
861
862
863


class TestRandomShortestSize:
    @pytest.mark.parametrize("min_size,max_size", [([5, 9], 20), ([5, 9], None)])
Philip Meier's avatar
Philip Meier committed
864
865
    def test__get_params(self, min_size, max_size):
        canvas_size = (3, 10)
866

867
        transform = transforms.RandomShortestSize(min_size=min_size, max_size=max_size, antialias=True)
868

Philip Meier's avatar
Philip Meier committed
869
        sample = make_image(canvas_size)
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
        params = transform._get_params([sample])

        assert "size" in params
        size = params["size"]

        assert isinstance(size, tuple) and len(size) == 2

        longer = max(size)
        shorter = min(size)
        if max_size is not None:
            assert longer <= max_size
            assert shorter <= max_size
        else:
            assert shorter in min_size


class TestLinearTransformation:
    def test_assertions(self):
        with pytest.raises(ValueError, match="transformation_matrix should be square"):
            transforms.LinearTransformation(torch.rand(2, 3), torch.rand(5))

        with pytest.raises(ValueError, match="mean_vector should have the same length"):
            transforms.LinearTransformation(torch.rand(3, 3), torch.rand(5))

    @pytest.mark.parametrize(
        "inpt",
        [
            122 * torch.ones(1, 3, 8, 8),
            122.0 * torch.ones(1, 3, 8, 8),
            datapoints.Image(122 * torch.ones(1, 3, 8, 8)),
            PIL.Image.new("RGB", (8, 8), (122, 122, 122)),
        ],
    )
    def test__transform(self, inpt):

        v = 121 * torch.ones(3 * 8 * 8)
        m = torch.ones(3 * 8 * 8, 3 * 8 * 8)
        transform = transforms.LinearTransformation(m, v)

        if isinstance(inpt, PIL.Image.Image):
910
            with pytest.raises(TypeError, match="does not support PIL images"):
911
912
913
914
915
916
917
918
919
920
921
922
923
                transform(inpt)
        else:
            output = transform(inpt)
            assert isinstance(output, torch.Tensor)
            assert output.unique() == 3 * 8 * 8
            assert output.dtype == inpt.dtype


class TestRandomResize:
    def test__get_params(self):
        min_size = 3
        max_size = 6

924
        transform = transforms.RandomResize(min_size=min_size, max_size=max_size, antialias=True)
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970

        for _ in range(10):
            params = transform._get_params([])

            assert isinstance(params["size"], list) and len(params["size"]) == 1
            size = params["size"][0]

            assert min_size <= size < max_size


class TestUniformTemporalSubsample:
    @pytest.mark.parametrize(
        "inpt",
        [
            torch.zeros(10, 3, 8, 8),
            torch.zeros(1, 10, 3, 8, 8),
            datapoints.Video(torch.zeros(1, 10, 3, 8, 8)),
        ],
    )
    def test__transform(self, inpt):
        num_samples = 5
        transform = transforms.UniformTemporalSubsample(num_samples)

        output = transform(inpt)
        assert type(output) is type(inpt)
        assert output.shape[-4] == num_samples
        assert output.dtype == inpt.dtype


# TODO: remove this test in 0.17 when the default of antialias changes to True
def test_antialias_warning():
    pil_img = PIL.Image.new("RGB", size=(10, 10), color=127)
    tensor_img = torch.randint(0, 256, size=(3, 10, 10), dtype=torch.uint8)
    tensor_video = torch.randint(0, 256, size=(2, 3, 10, 10), dtype=torch.uint8)

    match = "The default value of the antialias parameter"
    with pytest.warns(UserWarning, match=match):
        transforms.RandomResizedCrop((20, 20))(tensor_img)
    with pytest.warns(UserWarning, match=match):
        transforms.ScaleJitter((20, 20))(tensor_img)
    with pytest.warns(UserWarning, match=match):
        transforms.RandomShortestSize((20, 20))(tensor_img)
    with pytest.warns(UserWarning, match=match):
        transforms.RandomResize(10, 20)(tensor_img)

    with pytest.warns(UserWarning, match=match):
971
        F.resized_crop(datapoints.Image(tensor_img), 0, 0, 10, 10, (20, 20))
972
973

    with pytest.warns(UserWarning, match=match):
974
        F.resize(datapoints.Video(tensor_video), (20, 20))
975
    with pytest.warns(UserWarning, match=match):
976
        F.resized_crop(datapoints.Video(tensor_video), 0, 0, 10, 10, (20, 20))
977
978
979
980
981
982
983
984
985
986
987
988
989

    with warnings.catch_warnings():
        warnings.simplefilter("error")
        transforms.RandomResizedCrop((20, 20))(pil_img)
        transforms.ScaleJitter((20, 20))(pil_img)
        transforms.RandomShortestSize((20, 20))(pil_img)
        transforms.RandomResize(10, 20)(pil_img)

        transforms.RandomResizedCrop((20, 20), antialias=True)(tensor_img)
        transforms.ScaleJitter((20, 20), antialias=True)(tensor_img)
        transforms.RandomShortestSize((20, 20), antialias=True)(tensor_img)
        transforms.RandomResize(10, 20, antialias=True)(tensor_img)

990
991
        F.resized_crop(datapoints.Image(tensor_img), 0, 0, 10, 10, (20, 20), antialias=True)
        F.resized_crop(datapoints.Video(tensor_video), 0, 0, 10, 10, (20, 20), antialias=True)
992
993
994
995
996


@pytest.mark.parametrize("image_type", (PIL.Image, torch.Tensor, datapoints.Image))
@pytest.mark.parametrize("label_type", (torch.Tensor, int))
@pytest.mark.parametrize("dataset_return_type", (dict, tuple))
997
@pytest.mark.parametrize("to_tensor", (transforms.ToTensor, transforms.ToImage))
998
999
1000
1001
1002
1003
1004
def test_classif_preset(image_type, label_type, dataset_return_type, to_tensor):

    image = datapoints.Image(torch.randint(0, 256, size=(1, 3, 250, 250), dtype=torch.uint8))
    if image_type is PIL.Image:
        image = to_pil_image(image[0])
    elif image_type is torch.Tensor:
        image = image.as_subclass(torch.Tensor)
1005
        assert is_pure_tensor(image)
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016

    label = 1 if label_type is int else torch.tensor([1])

    if dataset_return_type is dict:
        sample = {
            "image": image,
            "label": label,
        }
    else:
        sample = image, label

1017
1018
1019
1020
1021
1022
    if to_tensor is transforms.ToTensor:
        with pytest.warns(UserWarning, match="deprecated and will be removed"):
            to_tensor = to_tensor()
    else:
        to_tensor = to_tensor()

1023
1024
    t = transforms.Compose(
        [
1025
            transforms.RandomResizedCrop((224, 224), antialias=True),
1026
1027
1028
1029
1030
            transforms.RandomHorizontalFlip(p=1),
            transforms.RandAugment(),
            transforms.TrivialAugmentWide(),
            transforms.AugMix(),
            transforms.AutoAugment(),
1031
            to_tensor,
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
            # TODO: ConvertImageDtype is a pass-through on PIL images, is that
            # intended?  This results in a failure if we convert to tensor after
            # it, because the image would still be uint8 which make Normalize
            # fail.
            transforms.ConvertImageDtype(torch.float),
            transforms.Normalize(mean=[0, 0, 0], std=[1, 1, 1]),
            transforms.RandomErasing(p=1),
        ]
    )

    out = t(sample)

    assert type(out) == type(sample)

    if dataset_return_type is tuple:
        out_image, out_label = out
    else:
        assert out.keys() == sample.keys()
        out_image, out_label = out.values()

    assert out_image.shape[-2:] == (224, 224)
    assert out_label == label


@pytest.mark.parametrize("image_type", (PIL.Image, torch.Tensor, datapoints.Image))
@pytest.mark.parametrize("data_augmentation", ("hflip", "lsj", "multiscale", "ssd", "ssdlite"))
1058
@pytest.mark.parametrize("to_tensor", (transforms.ToTensor, transforms.ToImage))
1059
1060
1061
@pytest.mark.parametrize("sanitize", (True, False))
def test_detection_preset(image_type, data_augmentation, to_tensor, sanitize):
    torch.manual_seed(0)
1062
1063
1064
1065
1066
1067
1068

    if to_tensor is transforms.ToTensor:
        with pytest.warns(UserWarning, match="deprecated and will be removed"):
            to_tensor = to_tensor()
    else:
        to_tensor = to_tensor()

1069
1070
1071
    if data_augmentation == "hflip":
        t = [
            transforms.RandomHorizontalFlip(p=1),
1072
            to_tensor,
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
            transforms.ConvertImageDtype(torch.float),
        ]
    elif data_augmentation == "lsj":
        t = [
            transforms.ScaleJitter(target_size=(1024, 1024), antialias=True),
            # Note: replaced FixedSizeCrop with RandomCrop, becuase we're
            # leaving FixedSizeCrop in prototype for now, and it expects Label
            # classes which we won't release yet.
            # transforms.FixedSizeCrop(
            #     size=(1024, 1024), fill=defaultdict(lambda: (123.0, 117.0, 104.0), {datapoints.Mask: 0})
            # ),
            transforms.RandomCrop((1024, 1024), pad_if_needed=True),
            transforms.RandomHorizontalFlip(p=1),
1086
            to_tensor,
1087
1088
1089
1090
1091
1092
1093
1094
            transforms.ConvertImageDtype(torch.float),
        ]
    elif data_augmentation == "multiscale":
        t = [
            transforms.RandomShortestSize(
                min_size=(480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800), max_size=1333, antialias=True
            ),
            transforms.RandomHorizontalFlip(p=1),
1095
            to_tensor,
1096
1097
1098
1099
1100
            transforms.ConvertImageDtype(torch.float),
        ]
    elif data_augmentation == "ssd":
        t = [
            transforms.RandomPhotometricDistort(p=1),
1101
            transforms.RandomZoomOut(fill={"others": (123.0, 117.0, 104.0), datapoints.Mask: 0}, p=1),
1102
1103
            transforms.RandomIoUCrop(),
            transforms.RandomHorizontalFlip(p=1),
1104
            to_tensor,
1105
1106
1107
1108
1109
1110
            transforms.ConvertImageDtype(torch.float),
        ]
    elif data_augmentation == "ssdlite":
        t = [
            transforms.RandomIoUCrop(),
            transforms.RandomHorizontalFlip(p=1),
1111
            to_tensor,
1112
1113
1114
            transforms.ConvertImageDtype(torch.float),
        ]
    if sanitize:
1115
        t += [transforms.SanitizeBoundingBoxes()]
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
    t = transforms.Compose(t)

    num_boxes = 5
    H = W = 250

    image = datapoints.Image(torch.randint(0, 256, size=(1, 3, H, W), dtype=torch.uint8))
    if image_type is PIL.Image:
        image = to_pil_image(image[0])
    elif image_type is torch.Tensor:
        image = image.as_subclass(torch.Tensor)
1126
        assert is_pure_tensor(image)
1127
1128
1129
1130
1131
1132

    label = torch.randint(0, 10, size=(num_boxes,))

    boxes = torch.randint(0, min(H, W) // 2, size=(num_boxes, 4))
    boxes[:, 2:] += boxes[:, :2]
    boxes = boxes.clamp(min=0, max=min(H, W))
Philip Meier's avatar
Philip Meier committed
1133
    boxes = datapoints.BoundingBoxes(boxes, format="XYXY", canvas_size=(H, W))
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145

    masks = datapoints.Mask(torch.randint(0, 2, size=(num_boxes, H, W), dtype=torch.uint8))

    sample = {
        "image": image,
        "label": label,
        "boxes": boxes,
        "masks": masks,
    }

    out = t(sample)

1146
    if isinstance(to_tensor, transforms.ToTensor) and image_type is not datapoints.Image:
1147
        assert is_pure_tensor(out["image"])
1148
1149
1150
1151
1152
1153
1154
1155
    else:
        assert isinstance(out["image"], datapoints.Image)
    assert isinstance(out["label"], type(sample["label"]))

    num_boxes_expected = {
        # ssd and ssdlite contain RandomIoUCrop which may "remove" some bbox. It
        # doesn't remove them strictly speaking, it just marks some boxes as
        # degenerate and those boxes will be later removed by
1156
        # SanitizeBoundingBoxes(), which we add to the pipelines if the sanitize
1157
1158
1159
        # param is True.
        # Note that the values below are probably specific to the random seed
        # set above (which is fine).
1160
        (True, "ssd"): 5,
1161
1162
1163
1164
1165
1166
1167
        (True, "ssdlite"): 4,
    }.get((sanitize, data_augmentation), num_boxes)

    assert out["boxes"].shape[0] == out["masks"].shape[0] == out["label"].shape[0] == num_boxes_expected


@pytest.mark.parametrize("min_size", (1, 10))
1168
@pytest.mark.parametrize("labels_getter", ("default", lambda inputs: inputs["labels"], None, lambda inputs: None))
1169
1170
1171
1172
1173
1174
1175
1176
@pytest.mark.parametrize("sample_type", (tuple, dict))
def test_sanitize_bounding_boxes(min_size, labels_getter, sample_type):

    if sample_type is tuple and not isinstance(labels_getter, str):
        # The "lambda inputs: inputs["labels"]" labels_getter used in this test
        # doesn't work if the input is a tuple.
        return

1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
    H, W = 256, 128

    boxes_and_validity = [
        ([0, 1, 10, 1], False),  # Y1 == Y2
        ([0, 1, 0, 20], False),  # X1 == X2
        ([0, 0, min_size - 1, 10], False),  # H < min_size
        ([0, 0, 10, min_size - 1], False),  # W < min_size
        ([0, 0, 10, H + 1], False),  # Y2 > H
        ([0, 0, W + 1, 10], False),  # X2 > W
        ([-1, 1, 10, 20], False),  # any < 0
        ([0, 0, -1, 20], False),  # any < 0
        ([0, 0, -10, -1], False),  # any < 0
        ([0, 0, min_size, 10], True),  # H < min_size
        ([0, 0, 10, min_size], True),  # W < min_size
        ([0, 0, W, H], True),  # TODO: Is that actually OK?? Should it be -1?
        ([1, 1, 30, 20], True),
        ([0, 0, 10, 10], True),
        ([1, 1, 30, 20], True),
    ]

    random.shuffle(boxes_and_validity)  # For test robustness: mix order of wrong and correct cases
    boxes, is_valid_mask = zip(*boxes_and_validity)
    valid_indices = [i for (i, is_valid) in enumerate(is_valid_mask) if is_valid]

    boxes = torch.tensor(boxes)
    labels = torch.arange(boxes.shape[0])

1204
    boxes = datapoints.BoundingBoxes(
1205
1206
        boxes,
        format=datapoints.BoundingBoxFormat.XYXY,
Philip Meier's avatar
Philip Meier committed
1207
        canvas_size=(H, W),
1208
1209
1210
    )

    masks = datapoints.Mask(torch.randint(0, 2, size=(boxes.shape[0], H, W)))
1211
1212
    whatever = torch.rand(10)
    input_img = torch.randint(0, 256, size=(1, 3, H, W), dtype=torch.uint8)
1213
    sample = {
1214
        "image": input_img,
1215
1216
        "labels": labels,
        "boxes": boxes,
1217
        "whatever": whatever,
1218
1219
1220
1221
        "None": None,
        "masks": masks,
    }

1222
1223
1224
1225
    if sample_type is tuple:
        img = sample.pop("image")
        sample = (img, sample)

1226
    out = transforms.SanitizeBoundingBoxes(min_size=min_size, labels_getter=labels_getter)(sample)
1227

1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
    if sample_type is tuple:
        out_image = out[0]
        out_labels = out[1]["labels"]
        out_boxes = out[1]["boxes"]
        out_masks = out[1]["masks"]
        out_whatever = out[1]["whatever"]
    else:
        out_image = out["image"]
        out_labels = out["labels"]
        out_boxes = out["boxes"]
        out_masks = out["masks"]
        out_whatever = out["whatever"]

    assert out_image is input_img
    assert out_whatever is whatever
1243

1244
    assert isinstance(out_boxes, datapoints.BoundingBoxes)
1245
1246
    assert isinstance(out_masks, datapoints.Mask)

1247
    if labels_getter is None or (callable(labels_getter) and labels_getter({"labels": "blah"}) is None):
1248
        assert out_labels is labels
1249
    else:
1250
1251
        assert isinstance(out_labels, torch.Tensor)
        assert out_boxes.shape[0] == out_labels.shape[0] == out_masks.shape[0]
1252
        # This works because we conveniently set labels to arange(num_boxes)
1253
        assert out_labels.tolist() == valid_indices
1254
1255
1256
1257


def test_sanitize_bounding_boxes_errors():

1258
    good_bbox = datapoints.BoundingBoxes(
1259
1260
        [[0, 0, 10, 10]],
        format=datapoints.BoundingBoxFormat.XYXY,
Philip Meier's avatar
Philip Meier committed
1261
        canvas_size=(20, 20),
1262
1263
1264
    )

    with pytest.raises(ValueError, match="min_size must be >= 1"):
1265
        transforms.SanitizeBoundingBoxes(min_size=0)
1266
    with pytest.raises(ValueError, match="labels_getter should either be 'default'"):
1267
        transforms.SanitizeBoundingBoxes(labels_getter=12)
1268
1269
1270

    with pytest.raises(ValueError, match="Could not infer where the labels are"):
        bad_labels_key = {"bbox": good_bbox, "BAD_KEY": torch.arange(good_bbox.shape[0])}
1271
        transforms.SanitizeBoundingBoxes()(bad_labels_key)
1272
1273
1274

    with pytest.raises(ValueError, match="must be a tensor"):
        not_a_tensor = {"bbox": good_bbox, "labels": torch.arange(good_bbox.shape[0]).tolist()}
1275
        transforms.SanitizeBoundingBoxes()(not_a_tensor)
1276
1277
1278

    with pytest.raises(ValueError, match="Number of boxes"):
        different_sizes = {"bbox": good_bbox, "labels": torch.arange(good_bbox.shape[0] + 3)}
1279
        transforms.SanitizeBoundingBoxes()(different_sizes)
1280

1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328

@pytest.mark.parametrize(
    "import_statement",
    (
        "from torchvision.transforms import v2",
        "import torchvision.transforms.v2",
        "from torchvision.transforms.v2 import Resize",
        "import torchvision.transforms.v2.functional",
        "from torchvision.transforms.v2.functional import resize",
        "from torchvision import datapoints",
        "from torchvision.datapoints import Image",
        "from torchvision.datasets import wrap_dataset_for_transforms_v2",
    ),
)
@pytest.mark.parametrize("call_disable_warning", (True, False))
def test_warnings_v2_namespaces(import_statement, call_disable_warning):
    if call_disable_warning:
        source = f"""
        import warnings
        import torchvision
        torchvision.disable_beta_transforms_warning()
        with warnings.catch_warnings():
            warnings.simplefilter("error")
            {import_statement}
        """
    else:
        source = f"""
        import pytest
        with pytest.warns(UserWarning, match="v2 namespaces are still Beta"):
            {import_statement}
        """
    assert_run_python_script(textwrap.dedent(source))


def test_no_warnings_v1_namespace():
    source = """
    import warnings
    with warnings.catch_warnings():
        warnings.simplefilter("error")
        import torchvision.transforms
        from torchvision import transforms
        import torchvision.transforms.functional
        from torchvision.transforms import Resize
        from torchvision.transforms.functional import resize
        from torchvision import datasets
        from torchvision.datasets import ImageNet
    """
    assert_run_python_script(textwrap.dedent(source))
Philip Meier's avatar
Philip Meier committed
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363


class TestLambda:
    inputs = pytest.mark.parametrize("input", [object(), torch.empty(()), np.empty(()), "string", 1, 0.0])

    @inputs
    def test_default(self, input):
        was_applied = False

        def was_applied_fn(input):
            nonlocal was_applied
            was_applied = True
            return input

        transform = transforms.Lambda(was_applied_fn)

        transform(input)

        assert was_applied

    @inputs
    def test_with_types(self, input):
        was_applied = False

        def was_applied_fn(input):
            nonlocal was_applied
            was_applied = True
            return input

        types = (torch.Tensor, np.ndarray)
        transform = transforms.Lambda(was_applied_fn, *types)

        transform(input)

        assert was_applied is isinstance(input, types)