SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Add DetrImageProcessorFast (#34063)

Browse Source 
* add fully functionning image_processing_detr_fast

* Create tensors on the correct device

* fix copies

* fix doc

* add tests equivalence cpu gpu

* fix doc en

* add relative imports and copied from

* Fix copies and nit
This commit is contained in:
Yoni Gozlan
2024-10-21 09:05:05 -04:00
committed by GitHub
parent 24bdc94da5
commit a4122813d1
16 changed files with 2779 additions and 1047 deletions
Download Patch File Download Diff File Expand all files Collapse all files

429
tests/models/conditional_detr/test_image_processing_conditional_detr.py
View File

@@ -282,96 +282,97 @@ class ConditionalDetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcess
images = [image_0, image_1]
annotations = [annotations_0, annotations_1]
image_processing = ConditionalDetrImageProcessor()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
for image_processing_class in self.image_processor_list:
image_processing = image_processing_class()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_batched_coco_panoptic_annotations with Detr->ConditionalDetr
def test_batched_coco_panoptic_annotations(self):
@@ -402,146 +403,148 @@ class ConditionalDetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcess
images = [image_0, image_1]
annotations = [annotation_0, annotation_1]
# encode them
image_processing = ConditionalDetrImageProcessor(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
for image_processing_class in self.image_processor_list:
# encode them
image_processing = image_processing_class(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_max_width_max_height_resizing_and_pad_strategy with Detr->ConditionalDetr
def test_max_width_max_height_resizing_and_pad_strategy(self):
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
for image_processing_class in self.image_processor_list:
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = ConditionalDetrImageProcessor(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = ConditionalDetrImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = ConditionalDetrImageProcessor(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = ConditionalDetrImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = ConditionalDetrImageProcessor(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = image_processing_class(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
def test_longest_edge_shortest_edge_resizing_strategy(self):
image_1 = torch.ones([958, 653, 3], dtype=torch.uint8)

429
tests/models/deformable_detr/test_image_processing_deformable_detr.py
View File

@@ -284,96 +284,97 @@ class DeformableDetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessi
images = [image_0, image_1]
annotations = [annotations_0, annotations_1]
image_processing = DeformableDetrImageProcessor()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
for image_processing_class in self.image_processor_list:
image_processing = image_processing_class()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_batched_coco_panoptic_annotations with Detr->DeformableDetr
def test_batched_coco_panoptic_annotations(self):
@@ -404,146 +405,148 @@ class DeformableDetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessi
images = [image_0, image_1]
annotations = [annotation_0, annotation_1]
# encode them
image_processing = DeformableDetrImageProcessor(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
for image_processing_class in self.image_processor_list:
# encode them
image_processing = image_processing_class(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_max_width_max_height_resizing_and_pad_strategy with Detr->DeformableDetr
def test_max_width_max_height_resizing_and_pad_strategy(self):
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
for image_processing_class in self.image_processor_list:
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = DeformableDetrImageProcessor(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = DeformableDetrImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = DeformableDetrImageProcessor(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = DeformableDetrImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = DeformableDetrImageProcessor(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = image_processing_class(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
def test_longest_edge_shortest_edge_resizing_strategy(self):
image_1 = torch.ones([958, 653, 3], dtype=torch.uint8)

865
tests/models/detr/test_image_processing_detr.py
View File

@@ -19,8 +19,8 @@ import unittest
import numpy as np
from transformers.testing_utils import require_torch, require_vision, slow
from transformers.utils import is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_torch_gpu, require_vision, slow
from transformers.utils import is_torch_available, is_torchvision_available, is_vision_available
from ...test_image_processing_common import AnnotationFormatTestMixin, ImageProcessingTestMixin, prepare_image_inputs
@@ -33,6 +33,9 @@ if is_vision_available():
from transformers import DetrImageProcessor
if is_torchvision_available():
from transformers import DetrImageProcessorFast
class DetrImageProcessingTester(unittest.TestCase):
def __init__(
@@ -51,6 +54,7 @@ class DetrImageProcessingTester(unittest.TestCase):
image_std=[0.5, 0.5, 0.5],
do_pad=True,
):
super().__init__()
# by setting size["longest_edge"] > max_resolution we're effectively not testing this :p
size = size if size is not None else {"shortest_edge": 18, "longest_edge": 1333}
self.parent = parent
@@ -132,6 +136,7 @@ class DetrImageProcessingTester(unittest.TestCase):
@require_vision
class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixin, unittest.TestCase):
image_processing_class = DetrImageProcessor if is_vision_available() else None
fast_image_processing_class = DetrImageProcessorFast if is_torchvision_available() else None
def setUp(self):
super().setUp()
@@ -142,26 +147,28 @@ class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixi
return self.image_processor_tester.prepare_image_processor_dict()
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "do_rescale"))
self.assertTrue(hasattr(image_processing, "rescale_factor"))
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_pad"))
for image_processing_class in self.image_processor_list:
image_processing = image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "do_rescale"))
self.assertTrue(hasattr(image_processing, "rescale_factor"))
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_pad"))
def test_image_processor_from_dict_with_kwargs(self):
image_processor = self.image_processing_class.from_dict(self.image_processor_dict)
self.assertEqual(image_processor.size, {"shortest_edge": 18, "longest_edge": 1333})
self.assertEqual(image_processor.do_pad, True)
for image_processing_class in self.image_processor_list:
image_processor = image_processing_class.from_dict(self.image_processor_dict)
self.assertEqual(image_processor.size, {"shortest_edge": 18, "longest_edge": 1333})
self.assertEqual(image_processor.do_pad, True)
image_processor = self.image_processing_class.from_dict(
self.image_processor_dict, size=42, max_size=84, pad_and_return_pixel_mask=False
)
self.assertEqual(image_processor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(image_processor.do_pad, False)
image_processor = image_processing_class.from_dict(
self.image_processor_dict, size=42, max_size=84, pad_and_return_pixel_mask=False
)
self.assertEqual(image_processor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(image_processor.do_pad, False)
def test_should_raise_if_annotation_format_invalid(self):
image_processor_dict = self.image_processor_tester.prepare_image_processor_dict()
@@ -178,12 +185,13 @@ class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixi
}
image_processor_params = {**image_processor_dict, **{"format": "_INVALID_FORMAT_"}}
image_processor = self.image_processing_class(**image_processor_params)
for image_processing_class in self.image_processor_list:
image_processor = image_processing_class(**image_processor_params)
with self.assertRaises(ValueError) as e:
image_processor(**params)
with self.assertRaises(ValueError) as e:
image_processor(**params)
self.assertTrue(str(e.exception).startswith("_INVALID_FORMAT_ is not a valid AnnotationFormat"))
self.assertTrue(str(e.exception).startswith("_INVALID_FORMAT_ is not a valid AnnotationFormat"))
def test_valid_coco_detection_annotations(self):
# prepare image and target
@@ -193,32 +201,33 @@ class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixi
params = {"image_id": 39769, "annotations": target}
# encode them
image_processing = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
for image_processing_class in self.image_processor_list:
# encode them
image_processing = image_processing_class.from_pretrained("facebook/detr-resnet-50")
# legal encodings (single image)
_ = image_processing(images=image, annotations=params, return_tensors="pt")
_ = image_processing(images=image, annotations=[params], return_tensors="pt")
# legal encodings (single image)
_ = image_processing(images=image, annotations=params, return_tensors="pt")
_ = image_processing(images=image, annotations=[params], return_tensors="pt")
# legal encodings (batch of one image)
_ = image_processing(images=[image], annotations=params, return_tensors="pt")
_ = image_processing(images=[image], annotations=[params], return_tensors="pt")
# legal encodings (batch of one image)
_ = image_processing(images=[image], annotations=params, return_tensors="pt")
_ = image_processing(images=[image], annotations=[params], return_tensors="pt")
# legal encoding (batch of more than one image)
n = 5
_ = image_processing(images=[image] * n, annotations=[params] * n, return_tensors="pt")
# legal encoding (batch of more than one image)
n = 5
_ = image_processing(images=[image] * n, annotations=[params] * n, return_tensors="pt")
# example of an illegal encoding (missing the 'image_id' key)
with self.assertRaises(ValueError) as e:
image_processing(images=image, annotations={"annotations": target}, return_tensors="pt")
# example of an illegal encoding (missing the 'image_id' key)
with self.assertRaises(ValueError) as e:
image_processing(images=image, annotations={"annotations": target}, return_tensors="pt")
self.assertTrue(str(e.exception).startswith("Invalid COCO detection annotations"))
self.assertTrue(str(e.exception).startswith("Invalid COCO detection annotations"))
# example of an illegal encoding (unequal lengths of images and annotations)
with self.assertRaises(ValueError) as e:
image_processing(images=[image] * n, annotations=[params] * (n - 1), return_tensors="pt")
# example of an illegal encoding (unequal lengths of images and annotations)
with self.assertRaises(ValueError) as e:
image_processing(images=[image] * n, annotations=[params] * (n - 1), return_tensors="pt")
self.assertTrue(str(e.exception) == "The number of images (5) and annotations (4) do not match.")
self.assertTrue(str(e.exception) == "The number of images (5) and annotations (4) do not match.")
@slow
def test_call_pytorch_with_coco_detection_annotations(self):
@@ -229,40 +238,41 @@ class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixi
target = {"image_id": 39769, "annotations": target}
# encode them
image_processing = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
encoding = image_processing(images=image, annotations=target, return_tensors="pt")
for image_processing_class in self.image_processor_list:
# encode them
image_processing = image_processing_class.from_pretrained("facebook/detr-resnet-50")
encoding = image_processing(images=image, annotations=target, return_tensors="pt")
# verify pixel values
expected_shape = torch.Size([1, 3, 800, 1066])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# verify pixel values
expected_shape = torch.Size([1, 3, 800, 1066])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
expected_slice = torch.tensor([0.2796, 0.3138, 0.3481])
self.assertTrue(torch.allclose(encoding["pixel_values"][0, 0, 0, :3], expected_slice, atol=1e-4))
expected_slice = torch.tensor([0.2796, 0.3138, 0.3481])
self.assertTrue(torch.allclose(encoding["pixel_values"][0, 0, 0, :3], expected_slice, atol=1e-4))
# verify area
expected_area = torch.tensor([5887.9600, 11250.2061, 489353.8438, 837122.7500, 147967.5156, 165732.3438])
self.assertTrue(torch.allclose(encoding["labels"][0]["area"], expected_area))
# verify boxes
expected_boxes_shape = torch.Size([6, 4])
self.assertEqual(encoding["labels"][0]["boxes"].shape, expected_boxes_shape)
expected_boxes_slice = torch.tensor([0.5503, 0.2765, 0.0604, 0.2215])
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"][0], expected_boxes_slice, atol=1e-3))
# verify image_id
expected_image_id = torch.tensor([39769])
self.assertTrue(torch.allclose(encoding["labels"][0]["image_id"], expected_image_id))
# verify is_crowd
expected_is_crowd = torch.tensor([0, 0, 0, 0, 0, 0])
self.assertTrue(torch.allclose(encoding["labels"][0]["iscrowd"], expected_is_crowd))
# verify class_labels
expected_class_labels = torch.tensor([75, 75, 63, 65, 17, 17])
self.assertTrue(torch.allclose(encoding["labels"][0]["class_labels"], expected_class_labels))
# verify orig_size
expected_orig_size = torch.tensor([480, 640])
self.assertTrue(torch.allclose(encoding["labels"][0]["orig_size"], expected_orig_size))
# verify size
expected_size = torch.tensor([800, 1066])
self.assertTrue(torch.allclose(encoding["labels"][0]["size"], expected_size))
# verify area
expected_area = torch.tensor([5887.9600, 11250.2061, 489353.8438, 837122.7500, 147967.5156, 165732.3438])
self.assertTrue(torch.allclose(encoding["labels"][0]["area"], expected_area))
# verify boxes
expected_boxes_shape = torch.Size([6, 4])
self.assertEqual(encoding["labels"][0]["boxes"].shape, expected_boxes_shape)
expected_boxes_slice = torch.tensor([0.5503, 0.2765, 0.0604, 0.2215])
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"][0], expected_boxes_slice, atol=1e-3))
# verify image_id
expected_image_id = torch.tensor([39769])
self.assertTrue(torch.allclose(encoding["labels"][0]["image_id"], expected_image_id))
# verify is_crowd
expected_is_crowd = torch.tensor([0, 0, 0, 0, 0, 0])
self.assertTrue(torch.allclose(encoding["labels"][0]["iscrowd"], expected_is_crowd))
# verify class_labels
expected_class_labels = torch.tensor([75, 75, 63, 65, 17, 17])
self.assertTrue(torch.allclose(encoding["labels"][0]["class_labels"], expected_class_labels))
# verify orig_size
expected_orig_size = torch.tensor([480, 640])
self.assertTrue(torch.allclose(encoding["labels"][0]["orig_size"], expected_orig_size))
# verify size
expected_size = torch.tensor([800, 1066])
self.assertTrue(torch.allclose(encoding["labels"][0]["size"], expected_size))
@slow
def test_call_pytorch_with_coco_panoptic_annotations(self):
@@ -275,43 +285,45 @@ class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixi
masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
# encode them
image_processing = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50-panoptic")
encoding = image_processing(images=image, annotations=target, masks_path=masks_path, return_tensors="pt")
for image_processing_class in self.image_processor_list:
# encode them
image_processing = image_processing_class.from_pretrained("facebook/detr-resnet-50-panoptic")
encoding = image_processing(images=image, annotations=target, masks_path=masks_path, return_tensors="pt")
# verify pixel values
expected_shape = torch.Size([1, 3, 800, 1066])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# verify pixel values
expected_shape = torch.Size([1, 3, 800, 1066])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
expected_slice = torch.tensor([0.2796, 0.3138, 0.3481])
self.assertTrue(torch.allclose(encoding["pixel_values"][0, 0, 0, :3], expected_slice, atol=1e-4))
expected_slice = torch.tensor([0.2796, 0.3138, 0.3481])
self.assertTrue(torch.allclose(encoding["pixel_values"][0, 0, 0, :3], expected_slice, atol=1e-4))
# verify area
expected_area = torch.tensor([147979.6875, 165527.0469, 484638.5938, 11292.9375, 5879.6562, 7634.1147])
self.assertTrue(torch.allclose(encoding["labels"][0]["area"], expected_area))
# verify boxes
expected_boxes_shape = torch.Size([6, 4])
self.assertEqual(encoding["labels"][0]["boxes"].shape, expected_boxes_shape)
expected_boxes_slice = torch.tensor([0.2625, 0.5437, 0.4688, 0.8625])
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"][0], expected_boxes_slice, atol=1e-3))
# verify image_id
expected_image_id = torch.tensor([39769])
self.assertTrue(torch.allclose(encoding["labels"][0]["image_id"], expected_image_id))
# verify is_crowd
expected_is_crowd = torch.tensor([0, 0, 0, 0, 0, 0])
self.assertTrue(torch.allclose(encoding["labels"][0]["iscrowd"], expected_is_crowd))
# verify class_labels
expected_class_labels = torch.tensor([17, 17, 63, 75, 75, 93])
self.assertTrue(torch.allclose(encoding["labels"][0]["class_labels"], expected_class_labels))
# verify masks
expected_masks_sum = 822873
self.assertEqual(encoding["labels"][0]["masks"].sum().item(), expected_masks_sum)
# verify orig_size
expected_orig_size = torch.tensor([480, 640])
self.assertTrue(torch.allclose(encoding["labels"][0]["orig_size"], expected_orig_size))
# verify size
expected_size = torch.tensor([800, 1066])
self.assertTrue(torch.allclose(encoding["labels"][0]["size"], expected_size))
# verify area
expected_area = torch.tensor([147979.6875, 165527.0469, 484638.5938, 11292.9375, 5879.6562, 7634.1147])
self.assertTrue(torch.allclose(encoding["labels"][0]["area"], expected_area))
# verify boxes
expected_boxes_shape = torch.Size([6, 4])
self.assertEqual(encoding["labels"][0]["boxes"].shape, expected_boxes_shape)
expected_boxes_slice = torch.tensor([0.2625, 0.5437, 0.4688, 0.8625])
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"][0], expected_boxes_slice, atol=1e-3))
# verify image_id
expected_image_id = torch.tensor([39769])
self.assertTrue(torch.allclose(encoding["labels"][0]["image_id"], expected_image_id))
# verify is_crowd
expected_is_crowd = torch.tensor([0, 0, 0, 0, 0, 0])
self.assertTrue(torch.allclose(encoding["labels"][0]["iscrowd"], expected_is_crowd))
# verify class_labels
expected_class_labels = torch.tensor([17, 17, 63, 75, 75, 93])
self.assertTrue(torch.allclose(encoding["labels"][0]["class_labels"], expected_class_labels))
# verify masks
expected_masks_sum = 822873
relative_error = torch.abs(encoding["labels"][0]["masks"].sum() - expected_masks_sum) / expected_masks_sum
self.assertTrue(relative_error < 1e-3)
# verify orig_size
expected_orig_size = torch.tensor([480, 640])
self.assertTrue(torch.allclose(encoding["labels"][0]["orig_size"], expected_orig_size))
# verify size
expected_size = torch.tensor([800, 1066])
self.assertTrue(torch.allclose(encoding["labels"][0]["size"], expected_size))
@slow
def test_batched_coco_detection_annotations(self):
@@ -340,96 +352,97 @@ class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixi
images = [image_0, image_1]
annotations = [annotations_0, annotations_1]
image_processing = DetrImageProcessor()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
for image_processing_class in self.image_processor_list:
image_processing = image_processing_class()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
def test_batched_coco_panoptic_annotations(self):
# prepare image, target and masks_path
@@ -459,194 +472,318 @@ class DetrImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixi
images = [image_0, image_1]
annotations = [annotation_0, annotation_1]
# encode them
image_processing = DetrImageProcessor(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
for image_processing_class in self.image_processor_list:
# encode them
image_processing = image_processing_class(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
def test_max_width_max_height_resizing_and_pad_strategy(self):
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
for image_processing_class in self.image_processor_list:
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = DetrImageProcessor(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = DetrImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = DetrImageProcessor(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = DetrImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = DetrImageProcessor(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = image_processing_class(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
def test_longest_edge_shortest_edge_resizing_strategy(self):
image_1 = torch.ones([958, 653, 3], dtype=torch.uint8)
for image_processing_class in self.image_processor_list:
image_1 = torch.ones([958, 653, 3], dtype=torch.uint8)
# max size is set; width < height;
# do_pad=False, longest_edge=640, shortest_edge=640, image=958x653 -> 640x436
image_processor = DetrImageProcessor(
size={"longest_edge": 640, "shortest_edge": 640},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 640, 436]))
# max size is set; width < height;
# do_pad=False, longest_edge=640, shortest_edge=640, image=958x653 -> 640x436
image_processor = image_processing_class(
size={"longest_edge": 640, "shortest_edge": 640},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 640, 436]))
image_2 = torch.ones([653, 958, 3], dtype=torch.uint8)
# max size is set; height < width;
# do_pad=False, longest_edge=640, shortest_edge=640, image=653x958 -> 436x640
image_processor = DetrImageProcessor(
size={"longest_edge": 640, "shortest_edge": 640},
do_pad=False,
)
inputs = image_processor(images=[image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 436, 640]))
image_2 = torch.ones([653, 958, 3], dtype=torch.uint8)
# max size is set; height < width;
# do_pad=False, longest_edge=640, shortest_edge=640, image=653x958 -> 436x640
image_processor = image_processing_class(
size={"longest_edge": 640, "shortest_edge": 640},
do_pad=False,
)
inputs = image_processor(images=[image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 436, 640]))
image_3 = torch.ones([100, 120, 3], dtype=torch.uint8)
# max size is set; width == size; height > max_size;
# do_pad=False, longest_edge=118, shortest_edge=100, image=120x100 -> 118x98
image_processor = DetrImageProcessor(
size={"longest_edge": 118, "shortest_edge": 100},
do_pad=False,
)
inputs = image_processor(images=[image_3], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 98, 118]))
image_3 = torch.ones([100, 120, 3], dtype=torch.uint8)
# max size is set; width == size; height > max_size;
# do_pad=False, longest_edge=118, shortest_edge=100, image=120x100 -> 118x98
image_processor = image_processing_class(
size={"longest_edge": 118, "shortest_edge": 100},
do_pad=False,
)
inputs = image_processor(images=[image_3], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 98, 118]))
image_4 = torch.ones([128, 50, 3], dtype=torch.uint8)
# max size is set; height == size; width < max_size;
# do_pad=False, longest_edge=256, shortest_edge=50, image=50x128 -> 50x128
image_processor = DetrImageProcessor(
size={"longest_edge": 256, "shortest_edge": 50},
do_pad=False,
)
inputs = image_processor(images=[image_4], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 128, 50]))
image_4 = torch.ones([128, 50, 3], dtype=torch.uint8)
# max size is set; height == size; width < max_size;
# do_pad=False, longest_edge=256, shortest_edge=50, image=50x128 -> 50x128
image_processor = image_processing_class(
size={"longest_edge": 256, "shortest_edge": 50},
do_pad=False,
)
inputs = image_processor(images=[image_4], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 128, 50]))
image_5 = torch.ones([50, 50, 3], dtype=torch.uint8)
# max size is set; height == width; width < max_size;
# do_pad=False, longest_edge=117, shortest_edge=50, image=50x50 -> 50x50
image_processor = DetrImageProcessor(
size={"longest_edge": 117, "shortest_edge": 50},
do_pad=False,
image_5 = torch.ones([50, 50, 3], dtype=torch.uint8)
# max size is set; height == width; width < max_size;
# do_pad=False, longest_edge=117, shortest_edge=50, image=50x50 -> 50x50
image_processor = image_processing_class(
size={"longest_edge": 117, "shortest_edge": 50},
do_pad=False,
)
inputs = image_processor(images=[image_5], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 50, 50]))
@slow
@require_torch_gpu
def test_fast_processor_equivalence_cpu_gpu_coco_detection_annotations(self):
# prepare image and target
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
with open("./tests/fixtures/tests_samples/COCO/coco_annotations.txt", "r") as f:
target = json.loads(f.read())
target = {"image_id": 39769, "annotations": target}
processor = self.image_processor_list[1].from_pretrained("facebook/detr-resnet-50")
# 1. run processor on CPU
encoding_cpu = processor(images=image, annotations=target, return_tensors="pt", device="cpu")
# 2. run processor on GPU
encoding_gpu = processor(images=image, annotations=target, return_tensors="pt", device="cuda")
# verify pixel values
self.assertEqual(encoding_cpu["pixel_values"].shape, encoding_gpu["pixel_values"].shape)
self.assertTrue(
torch.allclose(
encoding_cpu["pixel_values"][0, 0, 0, :3],
encoding_gpu["pixel_values"][0, 0, 0, :3].to("cpu"),
atol=1e-4,
)
)
inputs = image_processor(images=[image_5], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 50, 50]))
# verify area
self.assertTrue(torch.allclose(encoding_cpu["labels"][0]["area"], encoding_gpu["labels"][0]["area"].to("cpu")))
# verify boxes
self.assertEqual(encoding_cpu["labels"][0]["boxes"].shape, encoding_gpu["labels"][0]["boxes"].shape)
self.assertTrue(
torch.allclose(
encoding_cpu["labels"][0]["boxes"][0], encoding_gpu["labels"][0]["boxes"][0].to("cpu"), atol=1e-3
)
)
# verify image_id
self.assertTrue(
torch.allclose(encoding_cpu["labels"][0]["image_id"], encoding_gpu["labels"][0]["image_id"].to("cpu"))
)
# verify is_crowd
self.assertTrue(
torch.allclose(encoding_cpu["labels"][0]["iscrowd"], encoding_gpu["labels"][0]["iscrowd"].to("cpu"))
)
# verify class_labels
self.assertTrue(
torch.allclose(
encoding_cpu["labels"][0]["class_labels"], encoding_gpu["labels"][0]["class_labels"].to("cpu")
)
)
# verify orig_size
self.assertTrue(
torch.allclose(encoding_cpu["labels"][0]["orig_size"], encoding_gpu["labels"][0]["orig_size"].to("cpu"))
)
# verify size
self.assertTrue(torch.allclose(encoding_cpu["labels"][0]["size"], encoding_gpu["labels"][0]["size"].to("cpu")))
@slow
@require_torch_gpu
def test_fast_processor_equivalence_cpu_gpu_coco_panoptic_annotations(self):
# prepare image, target and masks_path
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
with open("./tests/fixtures/tests_samples/COCO/coco_panoptic_annotations.txt", "r") as f:
target = json.loads(f.read())
target = {"file_name": "000000039769.png", "image_id": 39769, "segments_info": target}
masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
processor = self.image_processor_list[1].from_pretrained("facebook/detr-resnet-50-panoptic")
# 1. run processor on CPU
encoding_cpu = processor(
images=image, annotations=target, masks_path=masks_path, return_tensors="pt", device="cpu"
)
# 2. run processor on GPU
encoding_gpu = processor(
images=image, annotations=target, masks_path=masks_path, return_tensors="pt", device="cuda"
)
# verify pixel values
self.assertEqual(encoding_cpu["pixel_values"].shape, encoding_gpu["pixel_values"].shape)
self.assertTrue(
torch.allclose(
encoding_cpu["pixel_values"][0, 0, 0, :3],
encoding_gpu["pixel_values"][0, 0, 0, :3].to("cpu"),
atol=1e-4,
)
)
# verify area
self.assertTrue(torch.allclose(encoding_cpu["labels"][0]["area"], encoding_gpu["labels"][0]["area"].to("cpu")))
# verify boxes
self.assertEqual(encoding_cpu["labels"][0]["boxes"].shape, encoding_gpu["labels"][0]["boxes"].shape)
self.assertTrue(
torch.allclose(
encoding_cpu["labels"][0]["boxes"][0], encoding_gpu["labels"][0]["boxes"][0].to("cpu"), atol=1e-3
)
)
# verify image_id
self.assertTrue(
torch.allclose(encoding_cpu["labels"][0]["image_id"], encoding_gpu["labels"][0]["image_id"].to("cpu"))
)
# verify is_crowd
self.assertTrue(
torch.allclose(encoding_cpu["labels"][0]["iscrowd"], encoding_gpu["labels"][0]["iscrowd"].to("cpu"))
)
# verify class_labels
self.assertTrue(
torch.allclose(
encoding_cpu["labels"][0]["class_labels"], encoding_gpu["labels"][0]["class_labels"].to("cpu")
)
)
# verify masks
masks_sum_cpu = encoding_cpu["labels"][0]["masks"].sum()
masks_sum_gpu = encoding_gpu["labels"][0]["masks"].sum()
relative_error = torch.abs(masks_sum_cpu - masks_sum_gpu) / masks_sum_cpu
self.assertTrue(relative_error < 1e-3)
# verify orig_size
self.assertTrue(
torch.allclose(encoding_cpu["labels"][0]["orig_size"], encoding_gpu["labels"][0]["orig_size"].to("cpu"))
)
# verify size
self.assertTrue(torch.allclose(encoding_cpu["labels"][0]["size"], encoding_gpu["labels"][0]["size"].to("cpu")))

429
tests/models/grounding_dino/test_image_processing_grounding_dino.py
View File

@@ -269,96 +269,97 @@ class GroundingDinoImageProcessingTest(AnnotationFormatTestMixin, ImageProcessin
images = [image_0, image_1]
annotations = [annotations_0, annotations_1]
image_processing = GroundingDinoImageProcessor()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
for image_processing_class in self.image_processor_list:
image_processing = image_processing_class()
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
return_tensors="pt", # do_convert_annotations=True
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.6879, 0.4609, 0.0755, 0.3691],
[0.2118, 0.3359, 0.2601, 0.1566],
[0.5011, 0.5000, 0.9979, 1.0000],
[0.5010, 0.5020, 0.9979, 0.9959],
[0.3284, 0.5944, 0.5884, 0.8112],
[0.8394, 0.5445, 0.3213, 0.9110],
]
)
expected_boxes_1 = torch.tensor(
[
[0.4130, 0.2765, 0.0453, 0.2215],
[0.1272, 0.2016, 0.1561, 0.0940],
[0.3757, 0.4933, 0.7488, 0.9865],
[0.3759, 0.5002, 0.7492, 0.9955],
[0.1971, 0.5456, 0.3532, 0.8646],
[0.5790, 0.4115, 0.3430, 0.7161],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
@slow
# Copied from tests.models.deformable_detr.test_image_processing_deformable_detr.DeformableDetrImageProcessingTest.test_call_pytorch_with_coco_panoptic_annotations with DeformableDetr->GroundingDino
@@ -440,146 +441,148 @@ class GroundingDinoImageProcessingTest(AnnotationFormatTestMixin, ImageProcessin
images = [image_0, image_1]
annotations = [annotation_0, annotation_1]
# encode them
image_processing = GroundingDinoImageProcessor(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
for image_processing_class in self.image_processor_list:
# encode them
image_processing = image_processing_class(format="coco_panoptic")
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_tensors="pt",
return_segmentation_masks=True,
)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the pixel values have been padded
postprocessed_height, postprocessed_width = 800, 1066
expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
self.assertEqual(encoding["pixel_values"].shape, expected_shape)
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the bounding boxes have been adjusted for padded images
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
expected_boxes_0 = torch.tensor(
[
[0.2625, 0.5437, 0.4688, 0.8625],
[0.7719, 0.4104, 0.4531, 0.7125],
[0.5000, 0.4927, 0.9969, 0.9854],
[0.1688, 0.2000, 0.2063, 0.0917],
[0.5492, 0.2760, 0.0578, 0.2187],
[0.4992, 0.4990, 0.9984, 0.9979],
]
)
expected_boxes_1 = torch.tensor(
[
[0.1576, 0.3262, 0.2814, 0.5175],
[0.4634, 0.2463, 0.2720, 0.4275],
[0.3002, 0.2956, 0.5985, 0.5913],
[0.1013, 0.1200, 0.1238, 0.0550],
[0.3297, 0.1656, 0.0347, 0.1312],
[0.2997, 0.2994, 0.5994, 0.5987],
]
)
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check the masks have also been padded
self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
# format and not in the range [0, 1]
encoding = image_processing(
images=images,
annotations=annotations,
masks_path=masks_path,
return_segmentation_masks=True,
do_convert_annotations=False,
return_tensors="pt",
)
self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
# Convert to absolute coordinates
unnormalized_boxes_0 = torch.vstack(
[
expected_boxes_0[:, 0] * postprocessed_width,
expected_boxes_0[:, 1] * postprocessed_height,
expected_boxes_0[:, 2] * postprocessed_width,
expected_boxes_0[:, 3] * postprocessed_height,
]
).T
unnormalized_boxes_1 = torch.vstack(
[
expected_boxes_1[:, 0] * postprocessed_width,
expected_boxes_1[:, 1] * postprocessed_height,
expected_boxes_1[:, 2] * postprocessed_width,
expected_boxes_1[:, 3] * postprocessed_height,
]
).T
# Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
expected_boxes_0 = torch.vstack(
[
unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
]
).T
expected_boxes_1 = torch.vstack(
[
unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
]
).T
self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
# Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_max_width_max_height_resizing_and_pad_strategy with Detr->GroundingDino
def test_max_width_max_height_resizing_and_pad_strategy(self):
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
for image_processing_class in self.image_processor_list:
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = GroundingDinoImageProcessor(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = GroundingDinoImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = GroundingDinoImageProcessor(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = GroundingDinoImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = GroundingDinoImageProcessor(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = image_processing_class(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
def test_longest_edge_shortest_edge_resizing_strategy(self):
image_1 = torch.ones([958, 653, 3], dtype=torch.uint8)

77
tests/models/yolos/test_image_processing_yolos.py
View File

@@ -553,47 +553,48 @@ class YolosImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMix
# Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_max_width_max_height_resizing_and_pad_strategy with Detr->Yolos
def test_max_width_max_height_resizing_and_pad_strategy(self):
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
for image_processing_class in self.image_processor_list:
image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = YolosImageProcessor(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = YolosImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=False,
)
inputs = image_processor(images=[image_1], return_tensors="pt")
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = YolosImageProcessor(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
image_processor = image_processing_class(
size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = YolosImageProcessor(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
# do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
image_processor = image_processing_class(
size={"max_height": 300, "max_width": 100},
do_pad=True,
pad_size={"height": 301, "width": 101},
)
inputs = image_processor(images=[image_1], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
### Check for batch
image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = YolosImageProcessor(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))
# do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
image_processor = image_processing_class(
size={"max_height": 150, "max_width": 100},
do_pad=True,
pad_size={"height": 150, "width": 100},
)
inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))