test_transforms_v2_refactored.py 86.3 KB
Newer Older
1
import contextlib
2
import decimal
3
import inspect
Philip Meier's avatar
Philip Meier committed
4
import math
5
6
7
8
9
10
11
12
13
14
15
16
17
18
import re
from unittest import mock

import numpy as np
import PIL.Image
import pytest

import torch
import torchvision.transforms.v2 as transforms
from common_utils import (
    assert_equal,
    assert_no_warnings,
    cache,
    cpu_and_cuda,
19
    freeze_rng_state,
20
21
22
23
    ignore_jit_no_profile_information_warning,
    make_bounding_box,
    make_detection_mask,
    make_image,
24
25
    make_image_pil,
    make_image_tensor,
26
27
    make_segmentation_mask,
    make_video,
28
    make_video_tensor,
29
    needs_cuda,
Nicolas Hug's avatar
Nicolas Hug committed
30
    set_rng_seed,
31
)
32
33

from torch import nn
34
from torch.testing import assert_close
35
from torch.utils._pytree import tree_map
36
from torch.utils.data import DataLoader, default_collate
37
from torchvision import datapoints
Philip Meier's avatar
Philip Meier committed
38
39

from torchvision.transforms._functional_tensor import _max_value as get_max_value
40
41
from torchvision.transforms.functional import pil_modes_mapping
from torchvision.transforms.v2 import functional as F
42
from torchvision.transforms.v2.functional._utils import _KERNEL_REGISTRY
43
44


Nicolas Hug's avatar
Nicolas Hug committed
45
46
47
48
49
50
@pytest.fixture(autouse=True)
def fix_rng_seed():
    set_rng_seed(0)
    yield


51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
def _to_tolerances(maybe_tolerance_dict):
    if not isinstance(maybe_tolerance_dict, dict):
        return dict(rtol=None, atol=None)

    tolerances = dict(rtol=0, atol=0)
    tolerances.update(maybe_tolerance_dict)
    return tolerances


def _check_kernel_cuda_vs_cpu(kernel, input, *args, rtol, atol, **kwargs):
    """Checks if the kernel produces closes results for inputs on GPU and CPU."""
    if input.device.type != "cuda":
        return

    input_cuda = input.as_subclass(torch.Tensor)
    input_cpu = input_cuda.to("cpu")

68
69
70
71
    with freeze_rng_state():
        actual = kernel(input_cuda, *args, **kwargs)
    with freeze_rng_state():
        expected = kernel(input_cpu, *args, **kwargs)
72
73
74
75
76

    assert_close(actual, expected, check_device=False, rtol=rtol, atol=atol)


@cache
77
def _script(obj):
78
    try:
79
        return torch.jit.script(obj)
80
    except Exception as error:
81
82
        name = getattr(obj, "__name__", obj.__class__.__name__)
        raise AssertionError(f"Trying to `torch.jit.script` '{name}' raised the error above.") from error
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138


def _check_kernel_scripted_vs_eager(kernel, input, *args, rtol, atol, **kwargs):
    """Checks if the kernel is scriptable and if the scripted output is close to the eager one."""
    if input.device.type != "cpu":
        return

    kernel_scripted = _script(kernel)

    input = input.as_subclass(torch.Tensor)
    with ignore_jit_no_profile_information_warning():
        actual = kernel_scripted(input, *args, **kwargs)
    expected = kernel(input, *args, **kwargs)

    assert_close(actual, expected, rtol=rtol, atol=atol)


def _check_kernel_batched_vs_unbatched(kernel, input, *args, rtol, atol, **kwargs):
    """Checks if the kernel produces close results for batched and unbatched inputs."""
    unbatched_input = input.as_subclass(torch.Tensor)

    for batch_dims in [(2,), (2, 1)]:
        repeats = [*batch_dims, *[1] * input.ndim]

        actual = kernel(unbatched_input.repeat(repeats), *args, **kwargs)

        expected = kernel(unbatched_input, *args, **kwargs)
        # We can't directly call `.repeat()` on the output, since some kernel also return some additional metadata
        if isinstance(expected, torch.Tensor):
            expected = expected.repeat(repeats)
        else:
            tensor, *metadata = expected
            expected = (tensor.repeat(repeats), *metadata)

        assert_close(actual, expected, rtol=rtol, atol=atol)

    for degenerate_batch_dims in [(0,), (5, 0), (0, 5)]:
        degenerate_batched_input = torch.empty(
            degenerate_batch_dims + input.shape, dtype=input.dtype, device=input.device
        )

        output = kernel(degenerate_batched_input, *args, **kwargs)
        # Most kernels just return a tensor, but some also return some additional metadata
        if not isinstance(output, torch.Tensor):
            output, *_ = output

        assert output.shape[: -input.ndim] == degenerate_batch_dims


def check_kernel(
    kernel,
    input,
    *args,
    check_cuda_vs_cpu=True,
    check_scripted_vs_eager=True,
    check_batched_vs_unbatched=True,
139
    expect_same_dtype=True,
140
141
142
143
144
145
146
147
148
149
150
151
    **kwargs,
):
    initial_input_version = input._version

    output = kernel(input.as_subclass(torch.Tensor), *args, **kwargs)
    # Most kernels just return a tensor, but some also return some additional metadata
    if not isinstance(output, torch.Tensor):
        output, *_ = output

    # check that no inplace operation happened
    assert input._version == initial_input_version

152
153
    if expect_same_dtype:
        assert output.dtype == input.dtype
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
    assert output.device == input.device

    if check_cuda_vs_cpu:
        _check_kernel_cuda_vs_cpu(kernel, input, *args, **kwargs, **_to_tolerances(check_cuda_vs_cpu))

    if check_scripted_vs_eager:
        _check_kernel_scripted_vs_eager(kernel, input, *args, **kwargs, **_to_tolerances(check_scripted_vs_eager))

    if check_batched_vs_unbatched:
        _check_kernel_batched_vs_unbatched(kernel, input, *args, **kwargs, **_to_tolerances(check_batched_vs_unbatched))


def _check_dispatcher_scripted_smoke(dispatcher, input, *args, **kwargs):
    """Checks if the dispatcher can be scripted and the scripted version can be called without error."""
    if not isinstance(input, datapoints.Image):
        return

    dispatcher_scripted = _script(dispatcher)
    with ignore_jit_no_profile_information_warning():
        dispatcher_scripted(input.as_subclass(torch.Tensor), *args, **kwargs)


def _check_dispatcher_dispatch(dispatcher, kernel, input, *args, **kwargs):
    """Checks if the dispatcher correctly dispatches the input to the corresponding kernel and that the input type is
    preserved in doing so. For bounding boxes also checks that the format is preserved.
    """
180
181
182
183
184
185
186
187
188
189
190
191
192
    input_type = type(input)

    if isinstance(input, datapoints.Datapoint):
        wrapped_kernel = _KERNEL_REGISTRY[dispatcher][input_type]

        # In case the wrapper was decorated with @functools.wraps, we can make the check more strict and test if the
        # proper kernel was wrapped
        if hasattr(wrapped_kernel, "__wrapped__"):
            assert wrapped_kernel.__wrapped__ is kernel

        spy = mock.MagicMock(wraps=wrapped_kernel, name=wrapped_kernel.__name__)
        with mock.patch.dict(_KERNEL_REGISTRY[dispatcher], values={input_type: spy}):
            output = dispatcher(input, *args, **kwargs)
193
194
195
196
197
198
199
200

        spy.assert_called_once()
    else:
        with mock.patch(f"{dispatcher.__module__}.{kernel.__name__}", wraps=kernel) as spy:
            output = dispatcher(input, *args, **kwargs)

            spy.assert_called_once()

201
    assert isinstance(output, input_type)
202

203
    if isinstance(input, datapoints.BoundingBoxes):
204
205
206
207
208
209
210
211
212
213
214
215
        assert output.format == input.format


def check_dispatcher(
    dispatcher,
    kernel,
    input,
    *args,
    check_scripted_smoke=True,
    check_dispatch=True,
    **kwargs,
):
216
    unknown_input = object()
217
    with mock.patch("torch._C._log_api_usage_once", wraps=torch._C._log_api_usage_once) as spy:
218
219
        with pytest.raises(TypeError, match=re.escape(str(type(unknown_input)))):
            dispatcher(unknown_input, *args, **kwargs)
220
221
222
223
224
225
226
227
228
229

        spy.assert_any_call(f"{dispatcher.__module__}.{dispatcher.__name__}")

    if check_scripted_smoke:
        _check_dispatcher_scripted_smoke(dispatcher, input, *args, **kwargs)

    if check_dispatch:
        _check_dispatcher_dispatch(dispatcher, kernel, input, *args, **kwargs)


230
def check_dispatcher_kernel_signature_match(dispatcher, *, kernel, input_type):
231
    """Checks if the signature of the dispatcher matches the kernel signature."""
232
233
    dispatcher_params = list(inspect.signature(dispatcher).parameters.values())[1:]
    kernel_params = list(inspect.signature(kernel).parameters.values())[1:]
234

235
    if issubclass(input_type, datapoints.Datapoint):
236
237
        # We filter out metadata that is implicitly passed to the dispatcher through the input datapoint, but has to be
        # explicitly passed to the kernel.
238
239
240
241
        explicit_metadata = {
            datapoints.BoundingBoxes: {"format", "canvas_size"},
        }
        kernel_params = [param for param in kernel_params if param.name not in explicit_metadata.get(input_type, set())]
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266

    dispatcher_params = iter(dispatcher_params)
    for dispatcher_param, kernel_param in zip(dispatcher_params, kernel_params):
        try:
            # In general, the dispatcher parameters are a superset of the kernel parameters. Thus, we filter out
            # dispatcher parameters that have no kernel equivalent while keeping the order intact.
            while dispatcher_param.name != kernel_param.name:
                dispatcher_param = next(dispatcher_params)
        except StopIteration:
            raise AssertionError(
                f"Parameter `{kernel_param.name}` of kernel `{kernel.__name__}` "
                f"has no corresponding parameter on the dispatcher `{dispatcher.__name__}`."
            ) from None

        if issubclass(input_type, PIL.Image.Image):
            # PIL kernels often have more correct annotations, since they are not limited by JIT. Thus, we don't check
            # them in the first place.
            dispatcher_param._annotation = kernel_param._annotation = inspect.Parameter.empty

        assert dispatcher_param == kernel_param


def _check_transform_v1_compatibility(transform, input):
    """If the transform defines the ``_v1_transform_cls`` attribute, checks if the transform has a public, static
    ``get_params`` method, is scriptable, and the scripted version can be called without error."""
267
    if transform._v1_transform_cls is None:
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
        return

    if type(input) is not torch.Tensor:
        return

    if hasattr(transform._v1_transform_cls, "get_params"):
        assert type(transform).get_params is transform._v1_transform_cls.get_params

    scripted_transform = _script(transform)
    with ignore_jit_no_profile_information_warning():
        scripted_transform(input)


def check_transform(transform_cls, input, *args, **kwargs):
    transform = transform_cls(*args, **kwargs)

    output = transform(input)
    assert isinstance(output, type(input))

287
    if isinstance(input, datapoints.BoundingBoxes):
288
289
290
291
292
        assert output.format == input.format

    _check_transform_v1_compatibility(transform, input)


293
def transform_cls_to_functional(transform_cls, **transform_specific_kwargs):
294
    def wrapper(input, *args, **kwargs):
295
        transform = transform_cls(*args, **transform_specific_kwargs, **kwargs)
296
297
298
299
300
301
302
        return transform(input)

    wrapper.__name__ = transform_cls.__name__

    return wrapper


Philip Meier's avatar
Philip Meier committed
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
def param_value_parametrization(**kwargs):
    """Helper function to turn

    @pytest.mark.parametrize(
        ("param", "value"),
        ("a", 1),
        ("a", 2),
        ("a", 3),
        ("b", -1.0)
        ("b", 1.0)
    )

    into

    @param_value_parametrization(a=[1, 2, 3], b=[-1.0, 1.0])
    """
    return pytest.mark.parametrize(
        ("param", "value"),
        [(param, value) for param, values in kwargs.items() for value in values],
    )


def adapt_fill(value, *, dtype):
    """Adapt fill values in the range [0.0, 1.0] to the value range of the dtype"""
    if value is None:
        return value

    max_value = get_max_value(dtype)

    if isinstance(value, (int, float)):
        return type(value)(value * max_value)
    elif isinstance(value, (list, tuple)):
        return type(value)(type(v)(v * max_value) for v in value)
    else:
        raise ValueError(f"fill should be an int or float, or a list or tuple of the former, but got '{value}'.")


EXHAUSTIVE_TYPE_FILLS = [
    None,
    1,
    0.5,
    [1],
    [0.2],
    (0,),
    (0.7,),
    [1, 0, 1],
    [0.1, 0.2, 0.3],
    (0, 1, 0),
    (0.9, 0.234, 0.314),
]
CORRECTNESS_FILLS = [
    v for v in EXHAUSTIVE_TYPE_FILLS if v is None or isinstance(v, float) or (isinstance(v, list) and len(v) > 1)
]


358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
# We cannot use `list(transforms.InterpolationMode)` here, since it includes some PIL-only ones as well
INTERPOLATION_MODES = [
    transforms.InterpolationMode.NEAREST,
    transforms.InterpolationMode.NEAREST_EXACT,
    transforms.InterpolationMode.BILINEAR,
    transforms.InterpolationMode.BICUBIC,
]


@contextlib.contextmanager
def assert_warns_antialias_default_value():
    with pytest.warns(UserWarning, match="The default value of the antialias parameter of all the resizing transforms"):
        yield


Philip Meier's avatar
Philip Meier committed
373
def reference_affine_bounding_boxes_helper(bounding_boxes, *, format, canvas_size, affine_matrix):
374
    def transform(bbox):
375
376
377
378
        # Go to float before converting to prevent precision loss in case of CXCYWH -> XYXY and W or H is 1
        in_dtype = bbox.dtype
        if not torch.is_floating_point(bbox):
            bbox = bbox.float()
379
        bbox_xyxy = F.convert_format_bounding_boxes(
380
            bbox.as_subclass(torch.Tensor),
381
            old_format=format,
382
383
384
385
386
387
388
389
390
391
392
            new_format=datapoints.BoundingBoxFormat.XYXY,
            inplace=True,
        )
        points = np.array(
            [
                [bbox_xyxy[0].item(), bbox_xyxy[1].item(), 1.0],
                [bbox_xyxy[2].item(), bbox_xyxy[1].item(), 1.0],
                [bbox_xyxy[0].item(), bbox_xyxy[3].item(), 1.0],
                [bbox_xyxy[2].item(), bbox_xyxy[3].item(), 1.0],
            ]
        )
393
        transformed_points = np.matmul(points, affine_matrix.T)
394
395
396
397
398
399
400
401
402
        out_bbox = torch.tensor(
            [
                np.min(transformed_points[:, 0]).item(),
                np.min(transformed_points[:, 1]).item(),
                np.max(transformed_points[:, 0]).item(),
                np.max(transformed_points[:, 1]).item(),
            ],
            dtype=bbox_xyxy.dtype,
        )
403
        out_bbox = F.convert_format_bounding_boxes(
404
            out_bbox, old_format=datapoints.BoundingBoxFormat.XYXY, new_format=format, inplace=True
405
406
        )
        # It is important to clamp before casting, especially for CXCYWH format, dtype=int64
Philip Meier's avatar
Philip Meier committed
407
        out_bbox = F.clamp_bounding_boxes(out_bbox, format=format, canvas_size=canvas_size)
408
409
410
        out_bbox = out_bbox.to(dtype=in_dtype)
        return out_bbox

411
    return torch.stack([transform(b) for b in bounding_boxes.reshape(-1, 4).unbind()]).reshape(bounding_boxes.shape)
412
413


414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
@pytest.mark.parametrize(
    ("dispatcher", "registered_datapoint_clss"),
    [(dispatcher, set(registry.keys())) for dispatcher, registry in _KERNEL_REGISTRY.items()],
)
def test_exhaustive_kernel_registration(dispatcher, registered_datapoint_clss):
    missing = {
        datapoints.Image,
        datapoints.BoundingBoxes,
        datapoints.Mask,
        datapoints.Video,
    } - registered_datapoint_clss
    if missing:
        names = sorted(f"datapoints.{cls.__name__}" for cls in missing)
        raise AssertionError(
            "\n".join(
                [
                    f"The dispatcher '{dispatcher.__name__}' has no kernel registered for",
                    "",
                    *[f"- {name}" for name in names],
                    "",
                    f"If available, register the kernels with @_register_kernel_internal({dispatcher.__name__}, ...).",
                    f"If not, register explicit no-ops with @_register_explicit_noop({', '.join(names)})",
                ]
            )
        )


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
468
469
470
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
class TestResize:
    INPUT_SIZE = (17, 11)
    OUTPUT_SIZES = [17, [17], (17,), [12, 13], (12, 13)]

    def _make_max_size_kwarg(self, *, use_max_size, size):
        if use_max_size:
            if not (isinstance(size, int) or len(size) == 1):
                # This would result in an `ValueError`
                return None

            max_size = (size if isinstance(size, int) else size[0]) + 1
        else:
            max_size = None

        return dict(max_size=max_size)

    def _compute_output_size(self, *, input_size, size, max_size):
        if not (isinstance(size, int) or len(size) == 1):
            return tuple(size)

        if not isinstance(size, int):
            size = size[0]

        old_height, old_width = input_size
        ratio = old_width / old_height
        if ratio > 1:
            new_height = size
            new_width = int(ratio * new_height)
        else:
            new_width = size
            new_height = int(new_width / ratio)

        if max_size is not None and max(new_height, new_width) > max_size:
            # Need to recompute the aspect ratio, since it might have changed due to rounding
            ratio = new_width / new_height
            if ratio > 1:
                new_width = max_size
                new_height = int(new_width / ratio)
            else:
                new_height = max_size
                new_width = int(new_height * ratio)

        return new_height, new_width

    @pytest.mark.parametrize("size", OUTPUT_SIZES)
    @pytest.mark.parametrize("interpolation", INTERPOLATION_MODES)
    @pytest.mark.parametrize("use_max_size", [True, False])
    @pytest.mark.parametrize("antialias", [True, False])
    @pytest.mark.parametrize("dtype", [torch.float32, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    def test_kernel_image_tensor(self, size, interpolation, use_max_size, antialias, dtype, device):
        if not (max_size_kwarg := self._make_max_size_kwarg(use_max_size=use_max_size, size=size)):
            return

        # In contrast to CPU, there is no native `InterpolationMode.BICUBIC` implementation for uint8 images on CUDA.
        # Internally, it uses the float path. Thus, we need to test with an enormous tolerance here to account for that.
        atol = 30 if transforms.InterpolationMode.BICUBIC and dtype is torch.uint8 else 1
        check_cuda_vs_cpu_tolerances = dict(rtol=0, atol=atol / 255 if dtype.is_floating_point else atol)

        check_kernel(
            F.resize_image_tensor,
502
            make_image(self.INPUT_SIZE, dtype=dtype, device=device),
503
504
505
506
507
508
509
510
511
512
513
514
515
            size=size,
            interpolation=interpolation,
            **max_size_kwarg,
            antialias=antialias,
            check_cuda_vs_cpu=check_cuda_vs_cpu_tolerances,
            check_scripted_vs_eager=not isinstance(size, int),
        )

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("size", OUTPUT_SIZES)
    @pytest.mark.parametrize("use_max_size", [True, False])
    @pytest.mark.parametrize("dtype", [torch.float32, torch.int64])
    @pytest.mark.parametrize("device", cpu_and_cuda())
516
    def test_kernel_bounding_boxes(self, format, size, use_max_size, dtype, device):
517
518
519
        if not (max_size_kwarg := self._make_max_size_kwarg(use_max_size=use_max_size, size=size)):
            return

520
        bounding_boxes = make_bounding_box(
521
            format=format,
Philip Meier's avatar
Philip Meier committed
522
            canvas_size=self.INPUT_SIZE,
523
524
            dtype=dtype,
            device=device,
Philip Meier's avatar
Philip Meier committed
525
        )
526
        check_kernel(
527
528
            F.resize_bounding_boxes,
            bounding_boxes,
Philip Meier's avatar
Philip Meier committed
529
            canvas_size=bounding_boxes.canvas_size,
530
531
532
533
534
            size=size,
            **max_size_kwarg,
            check_scripted_vs_eager=not isinstance(size, int),
        )

535
536
537
    @pytest.mark.parametrize("make_mask", [make_segmentation_mask, make_detection_mask])
    def test_kernel_mask(self, make_mask):
        check_kernel(F.resize_mask, make_mask(self.INPUT_SIZE), size=self.OUTPUT_SIZES[-1])
538
539

    def test_kernel_video(self):
540
        check_kernel(F.resize_video, make_video(self.INPUT_SIZE), size=self.OUTPUT_SIZES[-1], antialias=True)
541
542
543

    @pytest.mark.parametrize("size", OUTPUT_SIZES)
    @pytest.mark.parametrize(
544
        ("kernel", "make_input"),
545
        [
546
547
548
            (F.resize_image_tensor, make_image_tensor),
            (F.resize_image_pil, make_image_pil),
            (F.resize_image_tensor, make_image),
549
            (F.resize_bounding_boxes, make_bounding_box),
550
551
            (F.resize_mask, make_segmentation_mask),
            (F.resize_video, make_video),
552
553
        ],
    )
554
    def test_dispatcher(self, size, kernel, make_input):
555
556
557
        check_dispatcher(
            F.resize,
            kernel,
558
            make_input(self.INPUT_SIZE),
559
560
561
562
563
564
            size=size,
            antialias=True,
            check_scripted_smoke=not isinstance(size, int),
        )

    @pytest.mark.parametrize(
565
        ("kernel", "input_type"),
566
        [
567
568
569
            (F.resize_image_tensor, torch.Tensor),
            (F.resize_image_pil, PIL.Image.Image),
            (F.resize_image_tensor, datapoints.Image),
570
            (F.resize_bounding_boxes, datapoints.BoundingBoxes),
571
572
            (F.resize_mask, datapoints.Mask),
            (F.resize_video, datapoints.Video),
573
574
575
        ],
    )
    def test_dispatcher_signature(self, kernel, input_type):
576
        check_dispatcher_kernel_signature_match(F.resize, kernel=kernel, input_type=input_type)
577
578
579
580

    @pytest.mark.parametrize("size", OUTPUT_SIZES)
    @pytest.mark.parametrize("device", cpu_and_cuda())
    @pytest.mark.parametrize(
581
582
583
584
585
586
587
588
589
590
        "make_input",
        [
            make_image_tensor,
            make_image_pil,
            make_image,
            make_bounding_box,
            make_segmentation_mask,
            make_detection_mask,
            make_video,
        ],
591
    )
592
593
    def test_transform(self, size, device, make_input):
        check_transform(transforms.Resize, make_input(self.INPUT_SIZE, device=device), size=size, antialias=True)
594
595

    def _check_output_size(self, input, output, *, size, max_size):
Philip Meier's avatar
Philip Meier committed
596
597
        assert tuple(F.get_size(output)) == self._compute_output_size(
            input_size=F.get_size(input), size=size, max_size=max_size
598
599
600
601
602
603
604
605
606
607
608
609
        )

    @pytest.mark.parametrize("size", OUTPUT_SIZES)
    # `InterpolationMode.NEAREST` is modeled after the buggy `INTER_NEAREST` interpolation of CV2.
    # The PIL equivalent of `InterpolationMode.NEAREST` is `InterpolationMode.NEAREST_EXACT`
    @pytest.mark.parametrize("interpolation", set(INTERPOLATION_MODES) - {transforms.InterpolationMode.NEAREST})
    @pytest.mark.parametrize("use_max_size", [True, False])
    @pytest.mark.parametrize("fn", [F.resize, transform_cls_to_functional(transforms.Resize)])
    def test_image_correctness(self, size, interpolation, use_max_size, fn):
        if not (max_size_kwarg := self._make_max_size_kwarg(use_max_size=use_max_size, size=size)):
            return

610
        image = make_image(self.INPUT_SIZE, dtype=torch.uint8)
611
612
613
614
615
616
617
618
619

        actual = fn(image, size=size, interpolation=interpolation, **max_size_kwarg, antialias=True)
        expected = F.to_image_tensor(
            F.resize(F.to_image_pil(image), size=size, interpolation=interpolation, **max_size_kwarg)
        )

        self._check_output_size(image, actual, size=size, **max_size_kwarg)
        torch.testing.assert_close(actual, expected, atol=1, rtol=0)

620
    def _reference_resize_bounding_boxes(self, bounding_boxes, *, size, max_size=None):
Philip Meier's avatar
Philip Meier committed
621
        old_height, old_width = bounding_boxes.canvas_size
622
        new_height, new_width = self._compute_output_size(
Philip Meier's avatar
Philip Meier committed
623
            input_size=bounding_boxes.canvas_size, size=size, max_size=max_size
624
625
626
        )

        if (old_height, old_width) == (new_height, new_width):
627
            return bounding_boxes
628
629
630
631
632
633

        affine_matrix = np.array(
            [
                [new_width / old_width, 0, 0],
                [0, new_height / old_height, 0],
            ],
634
            dtype="float64" if bounding_boxes.dtype == torch.float64 else "float32",
635
636
        )

637
638
639
        expected_bboxes = reference_affine_bounding_boxes_helper(
            bounding_boxes,
            format=bounding_boxes.format,
Philip Meier's avatar
Philip Meier committed
640
            canvas_size=(new_height, new_width),
641
642
            affine_matrix=affine_matrix,
        )
Philip Meier's avatar
Philip Meier committed
643
        return datapoints.BoundingBoxes.wrap_like(bounding_boxes, expected_bboxes, canvas_size=(new_height, new_width))
644
645
646
647
648

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("size", OUTPUT_SIZES)
    @pytest.mark.parametrize("use_max_size", [True, False])
    @pytest.mark.parametrize("fn", [F.resize, transform_cls_to_functional(transforms.Resize)])
649
    def test_bounding_boxes_correctness(self, format, size, use_max_size, fn):
650
651
652
        if not (max_size_kwarg := self._make_max_size_kwarg(use_max_size=use_max_size, size=size)):
            return

Philip Meier's avatar
Philip Meier committed
653
        bounding_boxes = make_bounding_box(format=format, canvas_size=self.INPUT_SIZE)
654

655
656
        actual = fn(bounding_boxes, size=size, **max_size_kwarg)
        expected = self._reference_resize_bounding_boxes(bounding_boxes, size=size, **max_size_kwarg)
657

658
        self._check_output_size(bounding_boxes, actual, size=size, **max_size_kwarg)
659
660
661
662
        torch.testing.assert_close(actual, expected)

    @pytest.mark.parametrize("interpolation", set(transforms.InterpolationMode) - set(INTERPOLATION_MODES))
    @pytest.mark.parametrize(
663
664
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_video],
665
    )
666
667
    def test_pil_interpolation_compat_smoke(self, interpolation, make_input):
        input = make_input(self.INPUT_SIZE)
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682

        with (
            contextlib.nullcontext()
            if isinstance(input, PIL.Image.Image)
            # This error is triggered in PyTorch core
            else pytest.raises(NotImplementedError, match=f"got {interpolation.value.lower()}")
        ):
            F.resize(
                input,
                size=self.OUTPUT_SIZES[0],
                interpolation=interpolation,
            )

    def test_dispatcher_pil_antialias_warning(self):
        with pytest.warns(UserWarning, match="Anti-alias option is always applied for PIL Image input"):
683
            F.resize(make_image_pil(self.INPUT_SIZE), size=self.OUTPUT_SIZES[0], antialias=False)
684
685
686

    @pytest.mark.parametrize("size", OUTPUT_SIZES)
    @pytest.mark.parametrize(
687
688
689
690
691
692
693
694
695
696
        "make_input",
        [
            make_image_tensor,
            make_image_pil,
            make_image,
            make_bounding_box,
            make_segmentation_mask,
            make_detection_mask,
            make_video,
        ],
697
    )
698
    def test_max_size_error(self, size, make_input):
699
700
701
702
703
704
705
706
707
        if isinstance(size, int) or len(size) == 1:
            max_size = (size if isinstance(size, int) else size[0]) - 1
            match = "must be strictly greater than the requested size"
        else:
            # value can be anything other than None
            max_size = -1
            match = "size should be an int or a sequence of length 1"

        with pytest.raises(ValueError, match=match):
708
            F.resize(make_input(self.INPUT_SIZE), size=size, max_size=max_size, antialias=True)
709
710
711

    @pytest.mark.parametrize("interpolation", INTERPOLATION_MODES)
    @pytest.mark.parametrize(
712
713
        "make_input",
        [make_image_tensor, make_image, make_video],
714
    )
715
    def test_antialias_warning(self, interpolation, make_input):
716
717
718
719
720
        with (
            assert_warns_antialias_default_value()
            if interpolation in {transforms.InterpolationMode.BILINEAR, transforms.InterpolationMode.BICUBIC}
            else assert_no_warnings()
        ):
Philip Meier's avatar
Philip Meier committed
721
            F.resize(
722
                make_input(self.INPUT_SIZE),
Philip Meier's avatar
Philip Meier committed
723
724
725
                size=self.OUTPUT_SIZES[0],
                interpolation=interpolation,
            )
726
727
728

    @pytest.mark.parametrize("interpolation", INTERPOLATION_MODES)
    @pytest.mark.parametrize(
729
730
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_video],
731
    )
732
733
734
    def test_interpolation_int(self, interpolation, make_input):
        input = make_input(self.INPUT_SIZE)

735
736
737
        # `InterpolationMode.NEAREST_EXACT` has no proper corresponding integer equivalent. Internally, we map it to
        # `0` to be the same as `InterpolationMode.NEAREST` for PIL. However, for the tensor backend there is a
        # difference and thus we don't test it here.
738
        if isinstance(input, torch.Tensor) and interpolation is transforms.InterpolationMode.NEAREST_EXACT:
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
            return

        expected = F.resize(input, size=self.OUTPUT_SIZES[0], interpolation=interpolation, antialias=True)
        actual = F.resize(
            input, size=self.OUTPUT_SIZES[0], interpolation=pil_modes_mapping[interpolation], antialias=True
        )

        assert_equal(actual, expected)

    def test_transform_unknown_size_error(self):
        with pytest.raises(ValueError, match="size can either be an integer or a list or tuple of one or two integers"):
            transforms.Resize(size=object())

    @pytest.mark.parametrize(
        "size", [min(INPUT_SIZE), [min(INPUT_SIZE)], (min(INPUT_SIZE),), list(INPUT_SIZE), tuple(INPUT_SIZE)]
    )
    @pytest.mark.parametrize(
756
757
758
759
760
761
762
763
764
765
        "make_input",
        [
            make_image_tensor,
            make_image_pil,
            make_image,
            make_bounding_box,
            make_segmentation_mask,
            make_detection_mask,
            make_video,
        ],
766
    )
767
768
    def test_noop(self, size, make_input):
        input = make_input(self.INPUT_SIZE)
769

Philip Meier's avatar
Philip Meier committed
770
        output = F.resize(input, size=F.get_size(input), antialias=True)
771
772
773

        # This identity check is not a requirement. It is here to avoid breaking the behavior by accident. If there
        # is a good reason to break this, feel free to downgrade to an equality check.
774
        if isinstance(input, datapoints.Datapoint):
775
776
777
778
779
780
781
782
            # We can't test identity directly, since that checks for the identity of the Python object. Since all
            # datapoints unwrap before a kernel and wrap again afterwards, the Python object changes. Thus, we check
            # that the underlying storage is the same
            assert output.data_ptr() == input.data_ptr()
        else:
            assert output is input

    @pytest.mark.parametrize(
783
784
785
786
787
788
789
790
791
792
        "make_input",
        [
            make_image_tensor,
            make_image_pil,
            make_image,
            make_bounding_box,
            make_segmentation_mask,
            make_detection_mask,
            make_video,
        ],
793
    )
794
    def test_no_regression_5405(self, make_input):
795
796
797
        # Checks that `max_size` is not ignored if `size == small_edge_size`
        # See https://github.com/pytorch/vision/issues/5405

798
        input = make_input(self.INPUT_SIZE)
799

Philip Meier's avatar
Philip Meier committed
800
        size = min(F.get_size(input))
801
802
803
        max_size = size + 1
        output = F.resize(input, size=size, max_size=max_size, antialias=True)

Philip Meier's avatar
Philip Meier committed
804
        assert max(F.get_size(output)) == max_size
805
806
807
808
809
810


class TestHorizontalFlip:
    @pytest.mark.parametrize("dtype", [torch.float32, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    def test_kernel_image_tensor(self, dtype, device):
811
        check_kernel(F.horizontal_flip_image_tensor, make_image(dtype=dtype, device=device))
812
813
814
815

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("dtype", [torch.float32, torch.int64])
    @pytest.mark.parametrize("device", cpu_and_cuda())
816
817
    def test_kernel_bounding_boxes(self, format, dtype, device):
        bounding_boxes = make_bounding_box(format=format, dtype=dtype, device=device)
818
        check_kernel(
819
820
            F.horizontal_flip_bounding_boxes,
            bounding_boxes,
821
            format=format,
Philip Meier's avatar
Philip Meier committed
822
            canvas_size=bounding_boxes.canvas_size,
823
824
        )

825
826
827
    @pytest.mark.parametrize("make_mask", [make_segmentation_mask, make_detection_mask])
    def test_kernel_mask(self, make_mask):
        check_kernel(F.horizontal_flip_mask, make_mask())
828
829

    def test_kernel_video(self):
830
        check_kernel(F.horizontal_flip_video, make_video())
831
832

    @pytest.mark.parametrize(
833
        ("kernel", "make_input"),
834
        [
835
836
837
            (F.horizontal_flip_image_tensor, make_image_tensor),
            (F.horizontal_flip_image_pil, make_image_pil),
            (F.horizontal_flip_image_tensor, make_image),
838
            (F.horizontal_flip_bounding_boxes, make_bounding_box),
839
840
            (F.horizontal_flip_mask, make_segmentation_mask),
            (F.horizontal_flip_video, make_video),
841
842
        ],
    )
843
844
    def test_dispatcher(self, kernel, make_input):
        check_dispatcher(F.horizontal_flip, kernel, make_input())
845
846

    @pytest.mark.parametrize(
847
        ("kernel", "input_type"),
848
        [
849
850
851
            (F.horizontal_flip_image_tensor, torch.Tensor),
            (F.horizontal_flip_image_pil, PIL.Image.Image),
            (F.horizontal_flip_image_tensor, datapoints.Image),
852
            (F.horizontal_flip_bounding_boxes, datapoints.BoundingBoxes),
853
854
            (F.horizontal_flip_mask, datapoints.Mask),
            (F.horizontal_flip_video, datapoints.Video),
855
856
857
        ],
    )
    def test_dispatcher_signature(self, kernel, input_type):
858
        check_dispatcher_kernel_signature_match(F.horizontal_flip, kernel=kernel, input_type=input_type)
859
860

    @pytest.mark.parametrize(
861
862
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_bounding_box, make_segmentation_mask, make_video],
863
864
    )
    @pytest.mark.parametrize("device", cpu_and_cuda())
865
866
    def test_transform(self, make_input, device):
        check_transform(transforms.RandomHorizontalFlip, make_input(device=device), p=1)
867
868
869
870
871

    @pytest.mark.parametrize(
        "fn", [F.horizontal_flip, transform_cls_to_functional(transforms.RandomHorizontalFlip, p=1)]
    )
    def test_image_correctness(self, fn):
872
        image = make_image(dtype=torch.uint8, device="cpu")
873
874
875
876
877
878

        actual = fn(image)
        expected = F.to_image_tensor(F.horizontal_flip(F.to_image_pil(image)))

        torch.testing.assert_close(actual, expected)

879
    def _reference_horizontal_flip_bounding_boxes(self, bounding_boxes):
880
881
        affine_matrix = np.array(
            [
Philip Meier's avatar
Philip Meier committed
882
                [-1, 0, bounding_boxes.canvas_size[1]],
883
884
                [0, 1, 0],
            ],
885
            dtype="float64" if bounding_boxes.dtype == torch.float64 else "float32",
886
887
        )

888
889
890
        expected_bboxes = reference_affine_bounding_boxes_helper(
            bounding_boxes,
            format=bounding_boxes.format,
Philip Meier's avatar
Philip Meier committed
891
            canvas_size=bounding_boxes.canvas_size,
892
893
894
            affine_matrix=affine_matrix,
        )

895
        return datapoints.BoundingBoxes.wrap_like(bounding_boxes, expected_bboxes)
896
897
898
899
900

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize(
        "fn", [F.horizontal_flip, transform_cls_to_functional(transforms.RandomHorizontalFlip, p=1)]
    )
901
902
    def test_bounding_boxes_correctness(self, format, fn):
        bounding_boxes = make_bounding_box(format=format)
903

904
905
        actual = fn(bounding_boxes)
        expected = self._reference_horizontal_flip_bounding_boxes(bounding_boxes)
906
907
908
909

        torch.testing.assert_close(actual, expected)

    @pytest.mark.parametrize(
910
911
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_bounding_box, make_segmentation_mask, make_video],
912
913
    )
    @pytest.mark.parametrize("device", cpu_and_cuda())
914
915
    def test_transform_noop(self, make_input, device):
        input = make_input(device=device)
916
917
918
919
920
921

        transform = transforms.RandomHorizontalFlip(p=0)

        output = transform(input)

        assert_equal(output, input)
Philip Meier's avatar
Philip Meier committed
922
923
924
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


class TestAffine:
    _EXHAUSTIVE_TYPE_AFFINE_KWARGS = dict(
        # float, int
        angle=[-10.9, 18],
        # two-list of float, two-list of int, two-tuple of float, two-tuple of int
        translate=[[6.3, -0.6], [1, -3], (16.6, -6.6), (-2, 4)],
        # float
        scale=[0.5],
        # float, int,
        # one-list of float, one-list of int, one-tuple of float, one-tuple of int
        # two-list of float, two-list of int, two-tuple of float, two-tuple of int
        shear=[35.6, 38, [-37.7], [-23], (5.3,), (-52,), [5.4, 21.8], [-47, 51], (-11.2, 36.7), (8, -53)],
        # None
        # two-list of float, two-list of int, two-tuple of float, two-tuple of int
        center=[None, [1.2, 4.9], [-3, 1], (2.5, -4.7), (3, 2)],
    )
    # The special case for shear makes sure we pick a value that is supported while JIT scripting
    _MINIMAL_AFFINE_KWARGS = {
        k: vs[0] if k != "shear" else next(v for v in vs if isinstance(v, list))
        for k, vs in _EXHAUSTIVE_TYPE_AFFINE_KWARGS.items()
    }
    _CORRECTNESS_AFFINE_KWARGS = {
        k: [v for v in vs if v is None or isinstance(v, float) or (isinstance(v, list) and len(v) > 1)]
        for k, vs in _EXHAUSTIVE_TYPE_AFFINE_KWARGS.items()
    }

    _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES = dict(
        degrees=[30, (-15, 20)],
        translate=[None, (0.5, 0.5)],
        scale=[None, (0.75, 1.25)],
        shear=[None, (12, 30, -17, 5), 10, (-5, 12)],
    )
    _CORRECTNESS_TRANSFORM_AFFINE_RANGES = {
        k: next(v for v in vs if v is not None) for k, vs in _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES.items()
    }

    def _check_kernel(self, kernel, input, *args, **kwargs):
        kwargs_ = self._MINIMAL_AFFINE_KWARGS.copy()
        kwargs_.update(kwargs)
        check_kernel(kernel, input, *args, **kwargs_)

Philip Meier's avatar
Philip Meier committed
965
966
967
968
969
970
971
    @param_value_parametrization(
        angle=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["angle"],
        translate=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["translate"],
        shear=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["shear"],
        center=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["center"],
        interpolation=[transforms.InterpolationMode.NEAREST, transforms.InterpolationMode.BILINEAR],
        fill=EXHAUSTIVE_TYPE_FILLS,
Philip Meier's avatar
Philip Meier committed
972
973
974
975
976
    )
    @pytest.mark.parametrize("dtype", [torch.float32, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    def test_kernel_image_tensor(self, param, value, dtype, device):
        if param == "fill":
Philip Meier's avatar
Philip Meier committed
977
            value = adapt_fill(value, dtype=dtype)
Philip Meier's avatar
Philip Meier committed
978
979
        self._check_kernel(
            F.affine_image_tensor,
980
            make_image(dtype=dtype, device=device),
Philip Meier's avatar
Philip Meier committed
981
982
983
984
985
986
987
            **{param: value},
            check_scripted_vs_eager=not (param in {"shear", "fill"} and isinstance(value, (int, float))),
            check_cuda_vs_cpu=dict(atol=1, rtol=0)
            if dtype is torch.uint8 and param == "interpolation" and value is transforms.InterpolationMode.BILINEAR
            else True,
        )

Philip Meier's avatar
Philip Meier committed
988
989
990
991
992
    @param_value_parametrization(
        angle=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["angle"],
        translate=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["translate"],
        shear=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["shear"],
        center=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["center"],
Philip Meier's avatar
Philip Meier committed
993
994
995
996
    )
    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("dtype", [torch.float32, torch.int64])
    @pytest.mark.parametrize("device", cpu_and_cuda())
997
998
    def test_kernel_bounding_boxes(self, param, value, format, dtype, device):
        bounding_boxes = make_bounding_box(format=format, dtype=dtype, device=device)
Philip Meier's avatar
Philip Meier committed
999
        self._check_kernel(
1000
1001
            F.affine_bounding_boxes,
            bounding_boxes,
Philip Meier's avatar
Philip Meier committed
1002
            format=format,
Philip Meier's avatar
Philip Meier committed
1003
            canvas_size=bounding_boxes.canvas_size,
Philip Meier's avatar
Philip Meier committed
1004
1005
1006
1007
            **{param: value},
            check_scripted_vs_eager=not (param == "shear" and isinstance(value, (int, float))),
        )

1008
1009
1010
    @pytest.mark.parametrize("make_mask", [make_segmentation_mask, make_detection_mask])
    def test_kernel_mask(self, make_mask):
        self._check_kernel(F.affine_mask, make_mask())
Philip Meier's avatar
Philip Meier committed
1011
1012

    def test_kernel_video(self):
1013
        self._check_kernel(F.affine_video, make_video())
Philip Meier's avatar
Philip Meier committed
1014
1015

    @pytest.mark.parametrize(
1016
        ("kernel", "make_input"),
Philip Meier's avatar
Philip Meier committed
1017
        [
1018
1019
1020
            (F.affine_image_tensor, make_image_tensor),
            (F.affine_image_pil, make_image_pil),
            (F.affine_image_tensor, make_image),
1021
            (F.affine_bounding_boxes, make_bounding_box),
1022
1023
            (F.affine_mask, make_segmentation_mask),
            (F.affine_video, make_video),
Philip Meier's avatar
Philip Meier committed
1024
1025
        ],
    )
1026
1027
    def test_dispatcher(self, kernel, make_input):
        check_dispatcher(F.affine, kernel, make_input(), **self._MINIMAL_AFFINE_KWARGS)
Philip Meier's avatar
Philip Meier committed
1028
1029

    @pytest.mark.parametrize(
1030
        ("kernel", "input_type"),
Philip Meier's avatar
Philip Meier committed
1031
        [
1032
1033
1034
            (F.affine_image_tensor, torch.Tensor),
            (F.affine_image_pil, PIL.Image.Image),
            (F.affine_image_tensor, datapoints.Image),
1035
            (F.affine_bounding_boxes, datapoints.BoundingBoxes),
1036
1037
            (F.affine_mask, datapoints.Mask),
            (F.affine_video, datapoints.Video),
Philip Meier's avatar
Philip Meier committed
1038
1039
1040
        ],
    )
    def test_dispatcher_signature(self, kernel, input_type):
1041
        check_dispatcher_kernel_signature_match(F.affine, kernel=kernel, input_type=input_type)
Philip Meier's avatar
Philip Meier committed
1042
1043

    @pytest.mark.parametrize(
1044
1045
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_bounding_box, make_segmentation_mask, make_video],
Philip Meier's avatar
Philip Meier committed
1046
1047
    )
    @pytest.mark.parametrize("device", cpu_and_cuda())
1048
1049
    def test_transform(self, make_input, device):
        input = make_input(device=device)
Philip Meier's avatar
Philip Meier committed
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060

        check_transform(transforms.RandomAffine, input, **self._CORRECTNESS_TRANSFORM_AFFINE_RANGES)

    @pytest.mark.parametrize("angle", _CORRECTNESS_AFFINE_KWARGS["angle"])
    @pytest.mark.parametrize("translate", _CORRECTNESS_AFFINE_KWARGS["translate"])
    @pytest.mark.parametrize("scale", _CORRECTNESS_AFFINE_KWARGS["scale"])
    @pytest.mark.parametrize("shear", _CORRECTNESS_AFFINE_KWARGS["shear"])
    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
    @pytest.mark.parametrize(
        "interpolation", [transforms.InterpolationMode.NEAREST, transforms.InterpolationMode.BILINEAR]
    )
Philip Meier's avatar
Philip Meier committed
1061
    @pytest.mark.parametrize("fill", CORRECTNESS_FILLS)
Philip Meier's avatar
Philip Meier committed
1062
    def test_functional_image_correctness(self, angle, translate, scale, shear, center, interpolation, fill):
1063
        image = make_image(dtype=torch.uint8, device="cpu")
Philip Meier's avatar
Philip Meier committed
1064

Philip Meier's avatar
Philip Meier committed
1065
        fill = adapt_fill(fill, dtype=torch.uint8)
Philip Meier's avatar
Philip Meier committed
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096

        actual = F.affine(
            image,
            angle=angle,
            translate=translate,
            scale=scale,
            shear=shear,
            center=center,
            interpolation=interpolation,
            fill=fill,
        )
        expected = F.to_image_tensor(
            F.affine(
                F.to_image_pil(image),
                angle=angle,
                translate=translate,
                scale=scale,
                shear=shear,
                center=center,
                interpolation=interpolation,
                fill=fill,
            )
        )

        mae = (actual.float() - expected.float()).abs().mean()
        assert mae < 2 if interpolation is transforms.InterpolationMode.NEAREST else 8

    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
    @pytest.mark.parametrize(
        "interpolation", [transforms.InterpolationMode.NEAREST, transforms.InterpolationMode.BILINEAR]
    )
Philip Meier's avatar
Philip Meier committed
1097
    @pytest.mark.parametrize("fill", CORRECTNESS_FILLS)
Philip Meier's avatar
Philip Meier committed
1098
1099
    @pytest.mark.parametrize("seed", list(range(5)))
    def test_transform_image_correctness(self, center, interpolation, fill, seed):
1100
        image = make_image(dtype=torch.uint8, device="cpu")
Philip Meier's avatar
Philip Meier committed
1101

Philip Meier's avatar
Philip Meier committed
1102
        fill = adapt_fill(fill, dtype=torch.uint8)
Philip Meier's avatar
Philip Meier committed
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138

        transform = transforms.RandomAffine(
            **self._CORRECTNESS_TRANSFORM_AFFINE_RANGES, center=center, interpolation=interpolation, fill=fill
        )

        torch.manual_seed(seed)
        actual = transform(image)

        torch.manual_seed(seed)
        expected = F.to_image_tensor(transform(F.to_image_pil(image)))

        mae = (actual.float() - expected.float()).abs().mean()
        assert mae < 2 if interpolation is transforms.InterpolationMode.NEAREST else 8

    def _compute_affine_matrix(self, *, angle, translate, scale, shear, center):
        rot = math.radians(angle)
        cx, cy = center
        tx, ty = translate
        sx, sy = [math.radians(s) for s in ([shear, 0.0] if isinstance(shear, (int, float)) else shear)]

        c_matrix = np.array([[1, 0, cx], [0, 1, cy], [0, 0, 1]])
        t_matrix = np.array([[1, 0, tx], [0, 1, ty], [0, 0, 1]])
        c_matrix_inv = np.linalg.inv(c_matrix)
        rs_matrix = np.array(
            [
                [scale * math.cos(rot), -scale * math.sin(rot), 0],
                [scale * math.sin(rot), scale * math.cos(rot), 0],
                [0, 0, 1],
            ]
        )
        shear_x_matrix = np.array([[1, -math.tan(sx), 0], [0, 1, 0], [0, 0, 1]])
        shear_y_matrix = np.array([[1, 0, 0], [-math.tan(sy), 1, 0], [0, 0, 1]])
        rss_matrix = np.matmul(rs_matrix, np.matmul(shear_y_matrix, shear_x_matrix))
        true_matrix = np.matmul(t_matrix, np.matmul(c_matrix, np.matmul(rss_matrix, c_matrix_inv)))
        return true_matrix

1139
    def _reference_affine_bounding_boxes(self, bounding_boxes, *, angle, translate, scale, shear, center):
Philip Meier's avatar
Philip Meier committed
1140
        if center is None:
Philip Meier's avatar
Philip Meier committed
1141
            center = [s * 0.5 for s in bounding_boxes.canvas_size[::-1]]
Philip Meier's avatar
Philip Meier committed
1142
1143
1144
1145
1146
1147

        affine_matrix = self._compute_affine_matrix(
            angle=angle, translate=translate, scale=scale, shear=shear, center=center
        )
        affine_matrix = affine_matrix[:2, :]

1148
1149
1150
        expected_bboxes = reference_affine_bounding_boxes_helper(
            bounding_boxes,
            format=bounding_boxes.format,
Philip Meier's avatar
Philip Meier committed
1151
            canvas_size=bounding_boxes.canvas_size,
Philip Meier's avatar
Philip Meier committed
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
            affine_matrix=affine_matrix,
        )

        return expected_bboxes

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("angle", _CORRECTNESS_AFFINE_KWARGS["angle"])
    @pytest.mark.parametrize("translate", _CORRECTNESS_AFFINE_KWARGS["translate"])
    @pytest.mark.parametrize("scale", _CORRECTNESS_AFFINE_KWARGS["scale"])
    @pytest.mark.parametrize("shear", _CORRECTNESS_AFFINE_KWARGS["shear"])
    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
1163
1164
    def test_functional_bounding_boxes_correctness(self, format, angle, translate, scale, shear, center):
        bounding_boxes = make_bounding_box(format=format)
Philip Meier's avatar
Philip Meier committed
1165
1166

        actual = F.affine(
1167
            bounding_boxes,
Philip Meier's avatar
Philip Meier committed
1168
1169
1170
1171
1172
1173
            angle=angle,
            translate=translate,
            scale=scale,
            shear=shear,
            center=center,
        )
1174
1175
        expected = self._reference_affine_bounding_boxes(
            bounding_boxes,
Philip Meier's avatar
Philip Meier committed
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
            angle=angle,
            translate=translate,
            scale=scale,
            shear=shear,
            center=center,
        )

        torch.testing.assert_close(actual, expected)

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
    @pytest.mark.parametrize("seed", list(range(5)))
1188
1189
    def test_transform_bounding_boxes_correctness(self, format, center, seed):
        bounding_boxes = make_bounding_box(format=format)
Philip Meier's avatar
Philip Meier committed
1190
1191
1192
1193

        transform = transforms.RandomAffine(**self._CORRECTNESS_TRANSFORM_AFFINE_RANGES, center=center)

        torch.manual_seed(seed)
1194
        params = transform._get_params([bounding_boxes])
Philip Meier's avatar
Philip Meier committed
1195
1196

        torch.manual_seed(seed)
1197
        actual = transform(bounding_boxes)
Philip Meier's avatar
Philip Meier committed
1198

1199
        expected = self._reference_affine_bounding_boxes(bounding_boxes, **params, center=center)
Philip Meier's avatar
Philip Meier committed
1200
1201
1202
1203
1204
1205
1206
1207
1208

        torch.testing.assert_close(actual, expected)

    @pytest.mark.parametrize("degrees", _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES["degrees"])
    @pytest.mark.parametrize("translate", _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES["translate"])
    @pytest.mark.parametrize("scale", _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES["scale"])
    @pytest.mark.parametrize("shear", _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES["shear"])
    @pytest.mark.parametrize("seed", list(range(10)))
    def test_transform_get_params_bounds(self, degrees, translate, scale, shear, seed):
1209
        image = make_image()
Philip Meier's avatar
Philip Meier committed
1210
        height, width = F.get_size(image)
Philip Meier's avatar
Philip Meier committed
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283

        transform = transforms.RandomAffine(degrees=degrees, translate=translate, scale=scale, shear=shear)

        torch.manual_seed(seed)
        params = transform._get_params([image])

        if isinstance(degrees, (int, float)):
            assert -degrees <= params["angle"] <= degrees
        else:
            assert degrees[0] <= params["angle"] <= degrees[1]

        if translate is not None:
            width_max = int(round(translate[0] * width))
            height_max = int(round(translate[1] * height))
            assert -width_max <= params["translate"][0] <= width_max
            assert -height_max <= params["translate"][1] <= height_max
        else:
            assert params["translate"] == (0, 0)

        if scale is not None:
            assert scale[0] <= params["scale"] <= scale[1]
        else:
            assert params["scale"] == 1.0

        if shear is not None:
            if isinstance(shear, (int, float)):
                assert -shear <= params["shear"][0] <= shear
                assert params["shear"][1] == 0.0
            elif len(shear) == 2:
                assert shear[0] <= params["shear"][0] <= shear[1]
                assert params["shear"][1] == 0.0
            elif len(shear) == 4:
                assert shear[0] <= params["shear"][0] <= shear[1]
                assert shear[2] <= params["shear"][1] <= shear[3]
        else:
            assert params["shear"] == (0, 0)

    @pytest.mark.parametrize("param", ["degrees", "translate", "scale", "shear", "center"])
    @pytest.mark.parametrize("value", [0, [0], [0, 0, 0]])
    def test_transform_sequence_len_errors(self, param, value):
        if param in {"degrees", "shear"} and not isinstance(value, list):
            return

        kwargs = {param: value}
        if param != "degrees":
            kwargs["degrees"] = 0

        with pytest.raises(
            ValueError if isinstance(value, list) else TypeError, match=f"{param} should be a sequence of length 2"
        ):
            transforms.RandomAffine(**kwargs)

    def test_transform_negative_degrees_error(self):
        with pytest.raises(ValueError, match="If degrees is a single number, it must be positive"):
            transforms.RandomAffine(degrees=-1)

    @pytest.mark.parametrize("translate", [[-1, 0], [2, 0], [-1, 2]])
    def test_transform_translate_range_error(self, translate):
        with pytest.raises(ValueError, match="translation values should be between 0 and 1"):
            transforms.RandomAffine(degrees=0, translate=translate)

    @pytest.mark.parametrize("scale", [[-1, 0], [0, -1], [-1, -1]])
    def test_transform_scale_range_error(self, scale):
        with pytest.raises(ValueError, match="scale values should be positive"):
            transforms.RandomAffine(degrees=0, scale=scale)

    def test_transform_negative_shear_error(self):
        with pytest.raises(ValueError, match="If shear is a single number, it must be positive"):
            transforms.RandomAffine(degrees=0, shear=-1)

    def test_transform_unknown_fill_error(self):
        with pytest.raises(TypeError, match="Got inappropriate fill arg"):
            transforms.RandomAffine(degrees=0, fill="fill")
Philip Meier's avatar
Philip Meier committed
1284
1285
1286
1287
1288
1289


class TestVerticalFlip:
    @pytest.mark.parametrize("dtype", [torch.float32, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    def test_kernel_image_tensor(self, dtype, device):
1290
        check_kernel(F.vertical_flip_image_tensor, make_image(dtype=dtype, device=device))
Philip Meier's avatar
Philip Meier committed
1291
1292
1293
1294

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("dtype", [torch.float32, torch.int64])
    @pytest.mark.parametrize("device", cpu_and_cuda())
1295
1296
    def test_kernel_bounding_boxes(self, format, dtype, device):
        bounding_boxes = make_bounding_box(format=format, dtype=dtype, device=device)
Philip Meier's avatar
Philip Meier committed
1297
        check_kernel(
1298
1299
            F.vertical_flip_bounding_boxes,
            bounding_boxes,
Philip Meier's avatar
Philip Meier committed
1300
            format=format,
Philip Meier's avatar
Philip Meier committed
1301
            canvas_size=bounding_boxes.canvas_size,
Philip Meier's avatar
Philip Meier committed
1302
1303
        )

1304
1305
1306
    @pytest.mark.parametrize("make_mask", [make_segmentation_mask, make_detection_mask])
    def test_kernel_mask(self, make_mask):
        check_kernel(F.vertical_flip_mask, make_mask())
Philip Meier's avatar
Philip Meier committed
1307
1308

    def test_kernel_video(self):
1309
        check_kernel(F.vertical_flip_video, make_video())
Philip Meier's avatar
Philip Meier committed
1310
1311

    @pytest.mark.parametrize(
1312
        ("kernel", "make_input"),
Philip Meier's avatar
Philip Meier committed
1313
        [
1314
1315
1316
            (F.vertical_flip_image_tensor, make_image_tensor),
            (F.vertical_flip_image_pil, make_image_pil),
            (F.vertical_flip_image_tensor, make_image),
1317
            (F.vertical_flip_bounding_boxes, make_bounding_box),
1318
1319
            (F.vertical_flip_mask, make_segmentation_mask),
            (F.vertical_flip_video, make_video),
Philip Meier's avatar
Philip Meier committed
1320
1321
        ],
    )
1322
1323
    def test_dispatcher(self, kernel, make_input):
        check_dispatcher(F.vertical_flip, kernel, make_input())
Philip Meier's avatar
Philip Meier committed
1324
1325

    @pytest.mark.parametrize(
1326
        ("kernel", "input_type"),
Philip Meier's avatar
Philip Meier committed
1327
        [
1328
1329
1330
            (F.vertical_flip_image_tensor, torch.Tensor),
            (F.vertical_flip_image_pil, PIL.Image.Image),
            (F.vertical_flip_image_tensor, datapoints.Image),
1331
            (F.vertical_flip_bounding_boxes, datapoints.BoundingBoxes),
1332
1333
            (F.vertical_flip_mask, datapoints.Mask),
            (F.vertical_flip_video, datapoints.Video),
Philip Meier's avatar
Philip Meier committed
1334
1335
1336
        ],
    )
    def test_dispatcher_signature(self, kernel, input_type):
1337
        check_dispatcher_kernel_signature_match(F.vertical_flip, kernel=kernel, input_type=input_type)
Philip Meier's avatar
Philip Meier committed
1338
1339

    @pytest.mark.parametrize(
1340
1341
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_bounding_box, make_segmentation_mask, make_video],
Philip Meier's avatar
Philip Meier committed
1342
1343
    )
    @pytest.mark.parametrize("device", cpu_and_cuda())
1344
1345
    def test_transform(self, make_input, device):
        check_transform(transforms.RandomVerticalFlip, make_input(device=device), p=1)
Philip Meier's avatar
Philip Meier committed
1346
1347
1348

    @pytest.mark.parametrize("fn", [F.vertical_flip, transform_cls_to_functional(transforms.RandomVerticalFlip, p=1)])
    def test_image_correctness(self, fn):
1349
        image = make_image(dtype=torch.uint8, device="cpu")
Philip Meier's avatar
Philip Meier committed
1350
1351
1352
1353
1354
1355

        actual = fn(image)
        expected = F.to_image_tensor(F.vertical_flip(F.to_image_pil(image)))

        torch.testing.assert_close(actual, expected)

1356
    def _reference_vertical_flip_bounding_boxes(self, bounding_boxes):
Philip Meier's avatar
Philip Meier committed
1357
1358
1359
        affine_matrix = np.array(
            [
                [1, 0, 0],
Philip Meier's avatar
Philip Meier committed
1360
                [0, -1, bounding_boxes.canvas_size[0]],
Philip Meier's avatar
Philip Meier committed
1361
            ],
1362
            dtype="float64" if bounding_boxes.dtype == torch.float64 else "float32",
Philip Meier's avatar
Philip Meier committed
1363
1364
        )

1365
1366
1367
        expected_bboxes = reference_affine_bounding_boxes_helper(
            bounding_boxes,
            format=bounding_boxes.format,
Philip Meier's avatar
Philip Meier committed
1368
            canvas_size=bounding_boxes.canvas_size,
Philip Meier's avatar
Philip Meier committed
1369
1370
1371
            affine_matrix=affine_matrix,
        )

1372
        return datapoints.BoundingBoxes.wrap_like(bounding_boxes, expected_bboxes)
Philip Meier's avatar
Philip Meier committed
1373
1374
1375

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("fn", [F.vertical_flip, transform_cls_to_functional(transforms.RandomVerticalFlip, p=1)])
1376
1377
    def test_bounding_boxes_correctness(self, format, fn):
        bounding_boxes = make_bounding_box(format=format)
Philip Meier's avatar
Philip Meier committed
1378

1379
1380
        actual = fn(bounding_boxes)
        expected = self._reference_vertical_flip_bounding_boxes(bounding_boxes)
Philip Meier's avatar
Philip Meier committed
1381
1382
1383
1384

        torch.testing.assert_close(actual, expected)

    @pytest.mark.parametrize(
1385
1386
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_bounding_box, make_segmentation_mask, make_video],
Philip Meier's avatar
Philip Meier committed
1387
1388
    )
    @pytest.mark.parametrize("device", cpu_and_cuda())
1389
1390
    def test_transform_noop(self, make_input, device):
        input = make_input(device=device)
Philip Meier's avatar
Philip Meier committed
1391
1392
1393
1394
1395
1396

        transform = transforms.RandomVerticalFlip(p=0)

        output = transform(input)

        assert_equal(output, input)
Philip Meier's avatar
Philip Meier committed
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432


class TestRotate:
    _EXHAUSTIVE_TYPE_AFFINE_KWARGS = dict(
        # float, int
        angle=[-10.9, 18],
        # None
        # two-list of float, two-list of int, two-tuple of float, two-tuple of int
        center=[None, [1.2, 4.9], [-3, 1], (2.5, -4.7), (3, 2)],
    )
    _MINIMAL_AFFINE_KWARGS = {k: vs[0] for k, vs in _EXHAUSTIVE_TYPE_AFFINE_KWARGS.items()}
    _CORRECTNESS_AFFINE_KWARGS = {
        k: [v for v in vs if v is None or isinstance(v, float) or isinstance(v, list)]
        for k, vs in _EXHAUSTIVE_TYPE_AFFINE_KWARGS.items()
    }

    _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES = dict(
        degrees=[30, (-15, 20)],
    )
    _CORRECTNESS_TRANSFORM_AFFINE_RANGES = {k: vs[0] for k, vs in _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES.items()}

    @param_value_parametrization(
        angle=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["angle"],
        interpolation=[transforms.InterpolationMode.NEAREST, transforms.InterpolationMode.BILINEAR],
        expand=[False, True],
        center=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["center"],
        fill=EXHAUSTIVE_TYPE_FILLS,
    )
    @pytest.mark.parametrize("dtype", [torch.float32, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    def test_kernel_image_tensor(self, param, value, dtype, device):
        kwargs = {param: value}
        if param != "angle":
            kwargs["angle"] = self._MINIMAL_AFFINE_KWARGS["angle"]
        check_kernel(
            F.rotate_image_tensor,
1433
            make_image(dtype=dtype, device=device),
Philip Meier's avatar
Philip Meier committed
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
            **kwargs,
            check_scripted_vs_eager=not (param == "fill" and isinstance(value, (int, float))),
        )

    @param_value_parametrization(
        angle=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["angle"],
        expand=[False, True],
        center=_EXHAUSTIVE_TYPE_AFFINE_KWARGS["center"],
    )
    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("dtype", [torch.float32, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
1446
    def test_kernel_bounding_boxes(self, param, value, format, dtype, device):
Philip Meier's avatar
Philip Meier committed
1447
1448
1449
1450
        kwargs = {param: value}
        if param != "angle":
            kwargs["angle"] = self._MINIMAL_AFFINE_KWARGS["angle"]

1451
        bounding_boxes = make_bounding_box(format=format, dtype=dtype, device=device)
Philip Meier's avatar
Philip Meier committed
1452
1453

        check_kernel(
1454
1455
            F.rotate_bounding_boxes,
            bounding_boxes,
Philip Meier's avatar
Philip Meier committed
1456
            format=format,
Philip Meier's avatar
Philip Meier committed
1457
            canvas_size=bounding_boxes.canvas_size,
Philip Meier's avatar
Philip Meier committed
1458
1459
1460
            **kwargs,
        )

1461
1462
1463
    @pytest.mark.parametrize("make_mask", [make_segmentation_mask, make_detection_mask])
    def test_kernel_mask(self, make_mask):
        check_kernel(F.rotate_mask, make_mask(), **self._MINIMAL_AFFINE_KWARGS)
Philip Meier's avatar
Philip Meier committed
1464
1465

    def test_kernel_video(self):
1466
        check_kernel(F.rotate_video, make_video(), **self._MINIMAL_AFFINE_KWARGS)
Philip Meier's avatar
Philip Meier committed
1467
1468

    @pytest.mark.parametrize(
1469
        ("kernel", "make_input"),
Philip Meier's avatar
Philip Meier committed
1470
        [
1471
1472
1473
            (F.rotate_image_tensor, make_image_tensor),
            (F.rotate_image_pil, make_image_pil),
            (F.rotate_image_tensor, make_image),
1474
            (F.rotate_bounding_boxes, make_bounding_box),
1475
1476
            (F.rotate_mask, make_segmentation_mask),
            (F.rotate_video, make_video),
Philip Meier's avatar
Philip Meier committed
1477
1478
        ],
    )
1479
1480
    def test_dispatcher(self, kernel, make_input):
        check_dispatcher(F.rotate, kernel, make_input(), **self._MINIMAL_AFFINE_KWARGS)
Philip Meier's avatar
Philip Meier committed
1481
1482

    @pytest.mark.parametrize(
1483
        ("kernel", "input_type"),
Philip Meier's avatar
Philip Meier committed
1484
        [
1485
1486
1487
            (F.rotate_image_tensor, torch.Tensor),
            (F.rotate_image_pil, PIL.Image.Image),
            (F.rotate_image_tensor, datapoints.Image),
1488
            (F.rotate_bounding_boxes, datapoints.BoundingBoxes),
1489
1490
            (F.rotate_mask, datapoints.Mask),
            (F.rotate_video, datapoints.Video),
Philip Meier's avatar
Philip Meier committed
1491
1492
1493
        ],
    )
    def test_dispatcher_signature(self, kernel, input_type):
1494
        check_dispatcher_kernel_signature_match(F.rotate, kernel=kernel, input_type=input_type)
Philip Meier's avatar
Philip Meier committed
1495
1496

    @pytest.mark.parametrize(
1497
1498
        "make_input",
        [make_image_tensor, make_image_pil, make_image, make_bounding_box, make_segmentation_mask, make_video],
Philip Meier's avatar
Philip Meier committed
1499
1500
    )
    @pytest.mark.parametrize("device", cpu_and_cuda())
1501
1502
1503
1504
    def test_transform(self, make_input, device):
        check_transform(
            transforms.RandomRotation, make_input(device=device), **self._CORRECTNESS_TRANSFORM_AFFINE_RANGES
        )
Philip Meier's avatar
Philip Meier committed
1505
1506
1507
1508
1509
1510
1511
1512
1513

    @pytest.mark.parametrize("angle", _CORRECTNESS_AFFINE_KWARGS["angle"])
    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
    @pytest.mark.parametrize(
        "interpolation", [transforms.InterpolationMode.NEAREST, transforms.InterpolationMode.BILINEAR]
    )
    @pytest.mark.parametrize("expand", [False, True])
    @pytest.mark.parametrize("fill", CORRECTNESS_FILLS)
    def test_functional_image_correctness(self, angle, center, interpolation, expand, fill):
1514
        image = make_image(dtype=torch.uint8, device="cpu")
Philip Meier's avatar
Philip Meier committed
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535

        fill = adapt_fill(fill, dtype=torch.uint8)

        actual = F.rotate(image, angle=angle, center=center, interpolation=interpolation, expand=expand, fill=fill)
        expected = F.to_image_tensor(
            F.rotate(
                F.to_image_pil(image), angle=angle, center=center, interpolation=interpolation, expand=expand, fill=fill
            )
        )

        mae = (actual.float() - expected.float()).abs().mean()
        assert mae < 1 if interpolation is transforms.InterpolationMode.NEAREST else 6

    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
    @pytest.mark.parametrize(
        "interpolation", [transforms.InterpolationMode.NEAREST, transforms.InterpolationMode.BILINEAR]
    )
    @pytest.mark.parametrize("expand", [False, True])
    @pytest.mark.parametrize("fill", CORRECTNESS_FILLS)
    @pytest.mark.parametrize("seed", list(range(5)))
    def test_transform_image_correctness(self, center, interpolation, expand, fill, seed):
1536
        image = make_image(dtype=torch.uint8, device="cpu")
Philip Meier's avatar
Philip Meier committed
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556

        fill = adapt_fill(fill, dtype=torch.uint8)

        transform = transforms.RandomRotation(
            **self._CORRECTNESS_TRANSFORM_AFFINE_RANGES,
            center=center,
            interpolation=interpolation,
            expand=expand,
            fill=fill,
        )

        torch.manual_seed(seed)
        actual = transform(image)

        torch.manual_seed(seed)
        expected = F.to_image_tensor(transform(F.to_image_pil(image)))

        mae = (actual.float() - expected.float()).abs().mean()
        assert mae < 1 if interpolation is transforms.InterpolationMode.NEAREST else 6

1557
    def _reference_rotate_bounding_boxes(self, bounding_boxes, *, angle, expand, center):
Philip Meier's avatar
Philip Meier committed
1558
1559
1560
1561
1562
        # FIXME
        if expand:
            raise ValueError("This reference currently does not support expand=True")

        if center is None:
Philip Meier's avatar
Philip Meier committed
1563
            center = [s * 0.5 for s in bounding_boxes.canvas_size[::-1]]
Philip Meier's avatar
Philip Meier committed
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573

        a = np.cos(angle * np.pi / 180.0)
        b = np.sin(angle * np.pi / 180.0)
        cx = center[0]
        cy = center[1]
        affine_matrix = np.array(
            [
                [a, b, cx - cx * a - b * cy],
                [-b, a, cy + cx * b - a * cy],
            ],
1574
            dtype="float64" if bounding_boxes.dtype == torch.float64 else "float32",
Philip Meier's avatar
Philip Meier committed
1575
1576
        )

1577
1578
1579
        expected_bboxes = reference_affine_bounding_boxes_helper(
            bounding_boxes,
            format=bounding_boxes.format,
Philip Meier's avatar
Philip Meier committed
1580
            canvas_size=bounding_boxes.canvas_size,
Philip Meier's avatar
Philip Meier committed
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
            affine_matrix=affine_matrix,
        )

        return expected_bboxes

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    @pytest.mark.parametrize("angle", _CORRECTNESS_AFFINE_KWARGS["angle"])
    # TODO: add support for expand=True in the reference
    @pytest.mark.parametrize("expand", [False])
    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
1591
1592
    def test_functional_bounding_boxes_correctness(self, format, angle, expand, center):
        bounding_boxes = make_bounding_box(format=format)
Philip Meier's avatar
Philip Meier committed
1593

1594
1595
        actual = F.rotate(bounding_boxes, angle=angle, expand=expand, center=center)
        expected = self._reference_rotate_bounding_boxes(bounding_boxes, angle=angle, expand=expand, center=center)
Philip Meier's avatar
Philip Meier committed
1596
1597
1598
1599
1600
1601
1602
1603

        torch.testing.assert_close(actual, expected)

    @pytest.mark.parametrize("format", list(datapoints.BoundingBoxFormat))
    # TODO: add support for expand=True in the reference
    @pytest.mark.parametrize("expand", [False])
    @pytest.mark.parametrize("center", _CORRECTNESS_AFFINE_KWARGS["center"])
    @pytest.mark.parametrize("seed", list(range(5)))
1604
1605
    def test_transform_bounding_boxes_correctness(self, format, expand, center, seed):
        bounding_boxes = make_bounding_box(format=format)
Philip Meier's avatar
Philip Meier committed
1606
1607
1608
1609

        transform = transforms.RandomRotation(**self._CORRECTNESS_TRANSFORM_AFFINE_RANGES, expand=expand, center=center)

        torch.manual_seed(seed)
1610
        params = transform._get_params([bounding_boxes])
Philip Meier's avatar
Philip Meier committed
1611
1612

        torch.manual_seed(seed)
1613
        actual = transform(bounding_boxes)
Philip Meier's avatar
Philip Meier committed
1614

1615
        expected = self._reference_rotate_bounding_boxes(bounding_boxes, **params, expand=expand, center=center)
Philip Meier's avatar
Philip Meier committed
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653

        torch.testing.assert_close(actual, expected)

    @pytest.mark.parametrize("degrees", _EXHAUSTIVE_TYPE_TRANSFORM_AFFINE_RANGES["degrees"])
    @pytest.mark.parametrize("seed", list(range(10)))
    def test_transform_get_params_bounds(self, degrees, seed):
        transform = transforms.RandomRotation(degrees=degrees)

        torch.manual_seed(seed)
        params = transform._get_params([])

        if isinstance(degrees, (int, float)):
            assert -degrees <= params["angle"] <= degrees
        else:
            assert degrees[0] <= params["angle"] <= degrees[1]

    @pytest.mark.parametrize("param", ["degrees", "center"])
    @pytest.mark.parametrize("value", [0, [0], [0, 0, 0]])
    def test_transform_sequence_len_errors(self, param, value):
        if param == "degrees" and not isinstance(value, list):
            return

        kwargs = {param: value}
        if param != "degrees":
            kwargs["degrees"] = 0

        with pytest.raises(
            ValueError if isinstance(value, list) else TypeError, match=f"{param} should be a sequence of length 2"
        ):
            transforms.RandomRotation(**kwargs)

    def test_transform_negative_degrees_error(self):
        with pytest.raises(ValueError, match="If degrees is a single number, it must be positive"):
            transforms.RandomAffine(degrees=-1)

    def test_transform_unknown_fill_error(self):
        with pytest.raises(TypeError, match="Got inappropriate fill arg"):
            transforms.RandomAffine(degrees=0, fill="fill")
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714


class TestCompose:
    class BuiltinTransform(transforms.Transform):
        def _transform(self, inpt, params):
            return inpt

    class PackedInputTransform(nn.Module):
        def forward(self, sample):
            assert len(sample) == 2
            return sample

    class UnpackedInputTransform(nn.Module):
        def forward(self, image, label):
            return image, label

    @pytest.mark.parametrize(
        "transform_clss",
        [
            [BuiltinTransform],
            [PackedInputTransform],
            [UnpackedInputTransform],
            [BuiltinTransform, BuiltinTransform],
            [PackedInputTransform, PackedInputTransform],
            [UnpackedInputTransform, UnpackedInputTransform],
            [BuiltinTransform, PackedInputTransform, BuiltinTransform],
            [BuiltinTransform, UnpackedInputTransform, BuiltinTransform],
            [PackedInputTransform, BuiltinTransform, PackedInputTransform],
            [UnpackedInputTransform, BuiltinTransform, UnpackedInputTransform],
        ],
    )
    @pytest.mark.parametrize("unpack", [True, False])
    def test_packed_unpacked(self, transform_clss, unpack):
        needs_packed_inputs = any(issubclass(cls, self.PackedInputTransform) for cls in transform_clss)
        needs_unpacked_inputs = any(issubclass(cls, self.UnpackedInputTransform) for cls in transform_clss)
        assert not (needs_packed_inputs and needs_unpacked_inputs)

        transform = transforms.Compose([cls() for cls in transform_clss])

        image = make_image()
        label = 3
        packed_input = (image, label)

        def call_transform():
            if unpack:
                return transform(*packed_input)
            else:
                return transform(packed_input)

        if needs_unpacked_inputs and not unpack:
            with pytest.raises(TypeError, match="missing 1 required positional argument"):
                call_transform()
        elif needs_packed_inputs and unpack:
            with pytest.raises(TypeError, match="takes 2 positional arguments but 3 were given"):
                call_transform()
        else:
            output = call_transform()

            assert isinstance(output, tuple) and len(output) == 2
            assert output[0] is image
            assert output[1] is label
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841


class TestToDtype:
    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.to_dtype_image_tensor, make_image_tensor),
            (F.to_dtype_image_tensor, make_image),
            (F.to_dtype_video, make_video),
        ],
    )
    @pytest.mark.parametrize("input_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("output_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    @pytest.mark.parametrize("scale", (True, False))
    def test_kernel(self, kernel, make_input, input_dtype, output_dtype, device, scale):
        check_kernel(
            kernel,
            make_input(dtype=input_dtype, device=device),
            expect_same_dtype=input_dtype is output_dtype,
            dtype=output_dtype,
            scale=scale,
        )

    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.to_dtype_image_tensor, make_image_tensor),
            (F.to_dtype_image_tensor, make_image),
            (F.to_dtype_video, make_video),
        ],
    )
    @pytest.mark.parametrize("input_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("output_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    @pytest.mark.parametrize("scale", (True, False))
    def test_dispatcher(self, kernel, make_input, input_dtype, output_dtype, device, scale):
        check_dispatcher(
            F.to_dtype,
            kernel,
            make_input(dtype=input_dtype, device=device),
            # TODO: we could leave check_dispatch to True but it currently fails
            # in _check_dispatcher_dispatch because there is no to_dtype() method on the datapoints.
            # We should be able to put this back if we change the dispatch
            # mechanism e.g. via https://github.com/pytorch/vision/pull/7733
            check_dispatch=False,
            dtype=output_dtype,
            scale=scale,
        )

    @pytest.mark.parametrize(
        "make_input",
        [make_image_tensor, make_image, make_bounding_box, make_segmentation_mask, make_video],
    )
    @pytest.mark.parametrize("input_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("output_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    @pytest.mark.parametrize("scale", (True, False))
    @pytest.mark.parametrize("as_dict", (True, False))
    def test_transform(self, make_input, input_dtype, output_dtype, device, scale, as_dict):
        input = make_input(dtype=input_dtype, device=device)
        if as_dict:
            output_dtype = {type(input): output_dtype}
        check_transform(transforms.ToDtype, input, dtype=output_dtype, scale=scale)

    def reference_convert_dtype_image_tensor(self, image, dtype=torch.float, scale=False):
        input_dtype = image.dtype
        output_dtype = dtype

        if not scale:
            return image.to(dtype)

        if output_dtype == input_dtype:
            return image

        def fn(value):
            if input_dtype.is_floating_point:
                if output_dtype.is_floating_point:
                    return value
                else:
                    return round(decimal.Decimal(value) * torch.iinfo(output_dtype).max)
            else:
                input_max_value = torch.iinfo(input_dtype).max

                if output_dtype.is_floating_point:
                    return float(decimal.Decimal(value) / input_max_value)
                else:
                    output_max_value = torch.iinfo(output_dtype).max

                    if input_max_value > output_max_value:
                        factor = (input_max_value + 1) // (output_max_value + 1)
                        return value / factor
                    else:
                        factor = (output_max_value + 1) // (input_max_value + 1)
                        return value * factor

        return torch.tensor(tree_map(fn, image.tolist()), dtype=dtype, device=image.device)

    @pytest.mark.parametrize("input_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("output_dtype", [torch.float32, torch.float64, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    @pytest.mark.parametrize("scale", (True, False))
    def test_image_correctness(self, input_dtype, output_dtype, device, scale):
        if input_dtype.is_floating_point and output_dtype == torch.int64:
            pytest.xfail("float to int64 conversion is not supported")

        input = make_image(dtype=input_dtype, device=device)

        out = F.to_dtype(input, dtype=output_dtype, scale=scale)
        expected = self.reference_convert_dtype_image_tensor(input, dtype=output_dtype, scale=scale)

        if input_dtype.is_floating_point and not output_dtype.is_floating_point and scale:
            torch.testing.assert_close(out, expected, atol=1, rtol=0)
        else:
            torch.testing.assert_close(out, expected)

    def was_scaled(self, inpt):
        # this assumes the target dtype is float
        return inpt.max() <= 1

    def make_inpt_with_bbox_and_mask(self, make_input):
        H, W = 10, 10
        inpt_dtype = torch.uint8
        bbox_dtype = torch.float32
        mask_dtype = torch.bool
        sample = {
            "inpt": make_input(size=(H, W), dtype=inpt_dtype),
Philip Meier's avatar
Philip Meier committed
1842
            "bbox": make_bounding_box(canvas_size=(H, W), dtype=bbox_dtype),
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
            "mask": make_detection_mask(size=(H, W), dtype=mask_dtype),
        }

        return sample, inpt_dtype, bbox_dtype, mask_dtype

    @pytest.mark.parametrize("make_input", (make_image_tensor, make_image, make_video))
    @pytest.mark.parametrize("scale", (True, False))
    def test_dtype_not_a_dict(self, make_input, scale):
        # assert only inpt gets transformed when dtype isn't a dict

        sample, inpt_dtype, bbox_dtype, mask_dtype = self.make_inpt_with_bbox_and_mask(make_input)
        out = transforms.ToDtype(dtype=torch.float32, scale=scale)(sample)

        assert out["inpt"].dtype != inpt_dtype
        assert out["inpt"].dtype == torch.float32
        if scale:
            assert self.was_scaled(out["inpt"])
        else:
            assert not self.was_scaled(out["inpt"])
        assert out["bbox"].dtype == bbox_dtype
        assert out["mask"].dtype == mask_dtype

    @pytest.mark.parametrize("make_input", (make_image_tensor, make_image, make_video))
    def test_others_catch_all_and_none(self, make_input):
        # make sure "others" works as a catch-all and that None means no conversion

        sample, inpt_dtype, bbox_dtype, mask_dtype = self.make_inpt_with_bbox_and_mask(make_input)
        out = transforms.ToDtype(dtype={datapoints.Mask: torch.int64, "others": None})(sample)
        assert out["inpt"].dtype == inpt_dtype
        assert out["bbox"].dtype == bbox_dtype
        assert out["mask"].dtype != mask_dtype
        assert out["mask"].dtype == torch.int64

    @pytest.mark.parametrize("make_input", (make_image_tensor, make_image, make_video))
    def test_typical_use_case(self, make_input):
        # Typical use-case: want to convert dtype and scale for inpt and just dtype for masks.
        # This just makes sure we now have a decent API for this

        sample, inpt_dtype, bbox_dtype, mask_dtype = self.make_inpt_with_bbox_and_mask(make_input)
        out = transforms.ToDtype(
            dtype={type(sample["inpt"]): torch.float32, datapoints.Mask: torch.int64, "others": None}, scale=True
        )(sample)
        assert out["inpt"].dtype != inpt_dtype
        assert out["inpt"].dtype == torch.float32
        assert self.was_scaled(out["inpt"])
        assert out["bbox"].dtype == bbox_dtype
        assert out["mask"].dtype != mask_dtype
        assert out["mask"].dtype == torch.int64

    @pytest.mark.parametrize("make_input", (make_image_tensor, make_image, make_video))
    def test_errors_warnings(self, make_input):
        sample, inpt_dtype, bbox_dtype, mask_dtype = self.make_inpt_with_bbox_and_mask(make_input)

        with pytest.raises(ValueError, match="No dtype was specified for"):
            out = transforms.ToDtype(dtype={datapoints.Mask: torch.float32})(sample)
        with pytest.warns(UserWarning, match=re.escape("plain `torch.Tensor` will *not* be transformed")):
            transforms.ToDtype(dtype={torch.Tensor: torch.float32, datapoints.Image: torch.float32})
        with pytest.warns(UserWarning, match="no scaling will be done"):
            out = transforms.ToDtype(dtype={"others": None}, scale=True)(sample)
        assert out["inpt"].dtype == inpt_dtype
        assert out["bbox"].dtype == bbox_dtype
        assert out["mask"].dtype == mask_dtype
1905
1906


1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
class TestAdjustBrightness:
    _CORRECTNESS_BRIGHTNESS_FACTORS = [0.5, 0.0, 1.0, 5.0]
    _DEFAULT_BRIGHTNESS_FACTOR = _CORRECTNESS_BRIGHTNESS_FACTORS[0]

    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.adjust_brightness_image_tensor, make_image),
            (F.adjust_brightness_video, make_video),
        ],
    )
    @pytest.mark.parametrize("dtype", [torch.float32, torch.uint8])
    @pytest.mark.parametrize("device", cpu_and_cuda())
    def test_kernel(self, kernel, make_input, dtype, device):
        check_kernel(kernel, make_input(dtype=dtype, device=device), brightness_factor=self._DEFAULT_BRIGHTNESS_FACTOR)

    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.adjust_brightness_image_tensor, make_image_tensor),
            (F.adjust_brightness_image_pil, make_image_pil),
            (F.adjust_brightness_image_tensor, make_image),
            (F.adjust_brightness_video, make_video),
        ],
    )
    def test_dispatcher(self, kernel, make_input):
        check_dispatcher(F.adjust_brightness, kernel, make_input(), brightness_factor=self._DEFAULT_BRIGHTNESS_FACTOR)

    @pytest.mark.parametrize(
        ("kernel", "input_type"),
        [
            (F.adjust_brightness_image_tensor, torch.Tensor),
            (F.adjust_brightness_image_pil, PIL.Image.Image),
            (F.adjust_brightness_image_tensor, datapoints.Image),
            (F.adjust_brightness_video, datapoints.Video),
        ],
    )
    def test_dispatcher_signature(self, kernel, input_type):
        check_dispatcher_kernel_signature_match(F.adjust_brightness, kernel=kernel, input_type=input_type)

    @pytest.mark.parametrize("brightness_factor", _CORRECTNESS_BRIGHTNESS_FACTORS)
    def test_image_correctness(self, brightness_factor):
        image = make_image(dtype=torch.uint8, device="cpu")

        actual = F.adjust_brightness(image, brightness_factor=brightness_factor)
        expected = F.to_image_tensor(F.adjust_brightness(F.to_image_pil(image), brightness_factor=brightness_factor))

        torch.testing.assert_close(actual, expected)


1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
class TestCutMixMixUp:
    class DummyDataset:
        def __init__(self, size, num_classes):
            self.size = size
            self.num_classes = num_classes
            assert size < num_classes

        def __getitem__(self, idx):
            img = torch.rand(3, 100, 100)
            label = idx  # This ensures all labels in a batch are unique and makes testing easier
            return img, label

        def __len__(self):
            return self.size

Nicolas Hug's avatar
Nicolas Hug committed
1972
    @pytest.mark.parametrize("T", [transforms.CutMix, transforms.MixUp])
1973
1974
1975
1976
1977
1978
1979
    def test_supported_input_structure(self, T):

        batch_size = 32
        num_classes = 100

        dataset = self.DummyDataset(size=batch_size, num_classes=num_classes)

1980
        cutmix_mixup = T(num_classes=num_classes)
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021

        dl = DataLoader(dataset, batch_size=batch_size)

        # Input sanity checks
        img, target = next(iter(dl))
        input_img_size = img.shape[-3:]
        assert isinstance(img, torch.Tensor) and isinstance(target, torch.Tensor)
        assert target.shape == (batch_size,)

        def check_output(img, target):
            assert img.shape == (batch_size, *input_img_size)
            assert target.shape == (batch_size, num_classes)
            torch.testing.assert_close(target.sum(axis=-1), torch.ones(batch_size))
            num_non_zero_labels = (target != 0).sum(axis=-1)
            assert (num_non_zero_labels == 2).all()

        # After Dataloader, as unpacked input
        img, target = next(iter(dl))
        assert target.shape == (batch_size,)
        img, target = cutmix_mixup(img, target)
        check_output(img, target)

        # After Dataloader, as packed input
        packed_from_dl = next(iter(dl))
        assert isinstance(packed_from_dl, list)
        img, target = cutmix_mixup(packed_from_dl)
        check_output(img, target)

        # As collation function. We expect default_collate to be used by users.
        def collate_fn_1(batch):
            return cutmix_mixup(default_collate(batch))

        def collate_fn_2(batch):
            return cutmix_mixup(*default_collate(batch))

        for collate_fn in (collate_fn_1, collate_fn_2):
            dl = DataLoader(dataset, batch_size=batch_size, collate_fn=collate_fn)
            img, target = next(iter(dl))
            check_output(img, target)

    @needs_cuda
Nicolas Hug's avatar
Nicolas Hug committed
2022
    @pytest.mark.parametrize("T", [transforms.CutMix, transforms.MixUp])
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
    def test_cpu_vs_gpu(self, T):
        num_classes = 10
        batch_size = 3
        H, W = 12, 12

        imgs = torch.rand(batch_size, 3, H, W)
        labels = torch.randint(0, num_classes, (batch_size,))
        cutmix_mixup = T(alpha=0.5, num_classes=num_classes)

        _check_kernel_cuda_vs_cpu(cutmix_mixup, imgs, labels, rtol=None, atol=None)

Nicolas Hug's avatar
Nicolas Hug committed
2034
    @pytest.mark.parametrize("T", [transforms.CutMix, transforms.MixUp])
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
    def test_error(self, T):

        num_classes = 10
        batch_size = 9

        imgs = torch.rand(batch_size, 3, 12, 12)
        cutmix_mixup = T(alpha=0.5, num_classes=num_classes)

        for input_with_bad_type in (
            F.to_pil_image(imgs[0]),
            datapoints.Mask(torch.rand(12, 12)),
Philip Meier's avatar
Philip Meier committed
2046
            datapoints.BoundingBoxes(torch.rand(2, 4), format="XYXY", canvas_size=12),
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
        ):
            with pytest.raises(ValueError, match="does not support PIL images, "):
                cutmix_mixup(input_with_bad_type)

        with pytest.raises(ValueError, match="Could not infer where the labels are"):
            cutmix_mixup({"img": imgs, "Nothing_else": 3})

        with pytest.raises(ValueError, match="labels tensor should be of shape"):
            # Note: the error message isn't ideal, but that's because the label heuristic found the img as the label
            # It's OK, it's an edge-case. The important thing is that this fails loudly instead of passing silently
            cutmix_mixup(imgs)

        with pytest.raises(ValueError, match="When using the default labels_getter"):
            cutmix_mixup(imgs, "not_a_tensor")

        with pytest.raises(ValueError, match="labels tensor should be of shape"):
            cutmix_mixup(imgs, torch.randint(0, 2, size=(2, 3)))

        with pytest.raises(ValueError, match="Expected a batched input with 4 dims"):
            cutmix_mixup(imgs[None, None], torch.randint(0, num_classes, size=(batch_size,)))

        with pytest.raises(ValueError, match="does not match the batch size of the labels"):
            cutmix_mixup(imgs, torch.randint(0, num_classes, size=(batch_size + 1,)))

        with pytest.raises(ValueError, match="labels tensor should be of shape"):
            # The purpose of this check is more about documenting the current
            # behaviour of what happens on a Compose(), rather than actually
            # asserting the expected behaviour. We may support Compose() in the
            # future, e.g. for 2 consecutive CutMix?
            labels = torch.randint(0, num_classes, size=(batch_size,))
            transforms.Compose([cutmix_mixup, cutmix_mixup])(imgs, labels)


@pytest.mark.parametrize("key", ("labels", "LABELS", "LaBeL", "SOME_WEIRD_KEY_THAT_HAS_LABeL_IN_IT"))
@pytest.mark.parametrize("sample_type", (tuple, list, dict))
def test_labels_getter_default_heuristic(key, sample_type):
    labels = torch.arange(10)
    sample = {key: labels, "another_key": "whatever"}
    if sample_type is not dict:
        sample = sample_type((None, sample, "whatever_again"))
    assert transforms._utils._find_labels_default_heuristic(sample) is labels

    if key.lower() != "labels":
        # If "labels" is in the dict (case-insensitive),
        # it takes precedence over other keys which would otherwise be a match
        d = {key: "something_else", "labels": labels}
        assert transforms._utils._find_labels_default_heuristic(d) is labels
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183


class TestShapeGetters:
    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.get_dimensions_image_tensor, make_image_tensor),
            (F.get_dimensions_image_pil, make_image_pil),
            (F.get_dimensions_image_tensor, make_image),
            (F.get_dimensions_video, make_video),
        ],
    )
    def test_get_dimensions(self, kernel, make_input):
        size = (10, 10)
        color_space, num_channels = "RGB", 3

        input = make_input(size, color_space=color_space)

        assert kernel(input) == F.get_dimensions(input) == [num_channels, *size]

    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.get_num_channels_image_tensor, make_image_tensor),
            (F.get_num_channels_image_pil, make_image_pil),
            (F.get_num_channels_image_tensor, make_image),
            (F.get_num_channels_video, make_video),
        ],
    )
    def test_get_num_channels(self, kernel, make_input):
        color_space, num_channels = "RGB", 3

        input = make_input(color_space=color_space)

        assert kernel(input) == F.get_num_channels(input) == num_channels

    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.get_size_image_tensor, make_image_tensor),
            (F.get_size_image_pil, make_image_pil),
            (F.get_size_image_tensor, make_image),
            (F.get_size_bounding_boxes, make_bounding_box),
            (F.get_size_mask, make_detection_mask),
            (F.get_size_mask, make_segmentation_mask),
            (F.get_size_video, make_video),
        ],
    )
    def test_get_size(self, kernel, make_input):
        size = (10, 10)

        input = make_input(size)

        assert kernel(input) == F.get_size(input) == list(size)

    @pytest.mark.parametrize(
        ("kernel", "make_input"),
        [
            (F.get_num_frames_video, make_video_tensor),
            (F.get_num_frames_video, make_video),
        ],
    )
    def test_get_num_frames(self, kernel, make_input):
        num_frames = 4

        input = make_input(num_frames=num_frames)

        assert kernel(input) == F.get_num_frames(input) == num_frames

    @pytest.mark.parametrize(
        ("dispatcher", "make_input"),
        [
            (F.get_dimensions, make_bounding_box),
            (F.get_dimensions, make_detection_mask),
            (F.get_dimensions, make_segmentation_mask),
            (F.get_num_channels, make_bounding_box),
            (F.get_num_channels, make_detection_mask),
            (F.get_num_channels, make_segmentation_mask),
            (F.get_num_frames, make_image_pil),
            (F.get_num_frames, make_image),
            (F.get_num_frames, make_bounding_box),
            (F.get_num_frames, make_detection_mask),
            (F.get_num_frames, make_segmentation_mask),
        ],
    )
    def test_unsupported_types(self, dispatcher, make_input):
        input = make_input()

        with pytest.raises(TypeError, match=re.escape(str(type(input)))):
            dispatcher(input)