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

Update tests: replace feature extractor tests with image processor (#20768)

Browse Source 
* Update imports and test fetcher

* Revert but keep test fetcher update

* Fix imports

* Fix all imports

* Replace fe with ip names

* Add generate kwargs to `AutomaticSpeechRecognitionPipeline` (#20952)

* Add generate kwargs to AutomaticSpeechRecognitionPipeline

* Add test for generation kwargs

* Update image processor parameters if creating with kwargs (#20866)

* Update parameters if creating with kwargs

* Shallow copy to prevent mutating input

* Pass all args in constructor dict - warnings in init

* Fix typo

* Rename tester class

* Rebase and tidy up

* Fixup

* Use ImageProcessingSavingTestMixin

* Update property ref in tests

* Update property ref in tests

* Update recently merged in models

* Small fix

Co-authored-by: bofeng huang <bofenghuang7@gmail.com>
This commit is contained in:
amyeroberts
2023-01-23 17:25:41 +00:00
committed by GitHub
parent 354ea44340
commit e2bd7f80d0
31 changed files with 1974 additions and 2002 deletions
Download Patch File Download Diff File Expand all files Collapse all files

197
tests/models/beit/test_image_processing_beit.py
View File

@@ -22,8 +22,7 @@ from datasets import load_dataset
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -32,10 +31,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import BeitFeatureExtractor
from transformers import BeitImageProcessor
class BeitFeatureExtractionTester(unittest.TestCase):
class BeitImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -70,7 +69,7 @@ class BeitFeatureExtractionTester(unittest.TestCase):
self.image_std = image_std
self.do_reduce_labels = do_reduce_labels
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -105,166 +104,166 @@ def prepare_semantic_batch_inputs():
@require_torch
@require_vision
class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class BeitImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = BeitFeatureExtractor if is_vision_available() else None
image_processing_class = BeitImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = BeitFeatureExtractionTester(self)
self.image_processor_tester = BeitImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 20, "width": 20})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
self.assertEqual(feature_extractor.do_reduce_labels, False)
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, {"height": 20, "width": 20})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
self.assertEqual(image_processor.do_reduce_labels, False)
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, crop_size=84, reduce_labels=True
image_processor = self.image_processing_class.from_dict(
self.image_processor_dict, size=42, crop_size=84, reduce_labels=True
)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
self.assertEqual(feature_extractor.do_reduce_labels, True)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
self.assertEqual(image_processor.do_reduce_labels, True)
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_segmentation_maps(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
maps = []
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
maps.append(torch.zeros(image.shape[-2:]).long())
# Test not batched input
encoding = feature_extractor(image_inputs[0], maps[0], return_tensors="pt")
encoding = image_processing(image_inputs[0], maps[0], return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -272,22 +271,22 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
self.assertTrue(encoding["labels"].max().item() <= 255)
# Test batched
encoding = feature_extractor(image_inputs, maps, return_tensors="pt")
encoding = image_processing(image_inputs, maps, return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -297,22 +296,22 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
# Test not batched input (PIL images)
image, segmentation_map = prepare_semantic_single_inputs()
encoding = feature_extractor(image, segmentation_map, return_tensors="pt")
encoding = image_processing(image, segmentation_map, return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -322,22 +321,22 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
# Test batched input (PIL images)
images, segmentation_maps = prepare_semantic_batch_inputs()
encoding = feature_extractor(images, segmentation_maps, return_tensors="pt")
encoding = image_processing(images, segmentation_maps, return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
2,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
2,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -345,16 +344,16 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
self.assertTrue(encoding["labels"].max().item() <= 255)
def test_reduce_labels(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# ADE20k has 150 classes, and the background is included, so labels should be between 0 and 150
image, map = prepare_semantic_single_inputs()
encoding = feature_extractor(image, map, return_tensors="pt")
encoding = image_processing(image, map, return_tensors="pt")
self.assertTrue(encoding["labels"].min().item() >= 0)
self.assertTrue(encoding["labels"].max().item() <= 150)
feature_extractor.reduce_labels = True
encoding = feature_extractor(image, map, return_tensors="pt")
image_processing.reduce_labels = True
encoding = image_processing(image, map, return_tensors="pt")
self.assertTrue(encoding["labels"].min().item() >= 0)
self.assertTrue(encoding["labels"].max().item() <= 255)

148
tests/models/blip/test_image_processing_blip.py
View File

@@ -21,7 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import ImageProcessingSavingTestMixin
if is_torch_available():
@@ -65,7 +65,7 @@ class BlipImageProcessingTester(unittest.TestCase):
self.do_pad = do_pad
self.do_convert_rgb = do_convert_rgb
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -109,180 +109,180 @@ class BlipImageProcessingTester(unittest.TestCase):
@require_torch
@require_vision
class BlipImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class BlipImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = BlipImageProcessor if is_vision_available() else None
image_processing_class = BlipImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = BlipImageProcessingTester(self)
self.image_processor_tester = BlipImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_convert_rgb"))
def test_image_processor_properties(self):
image_processor = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processor, "do_resize"))
self.assertTrue(hasattr(image_processor, "size"))
self.assertTrue(hasattr(image_processor, "do_normalize"))
self.assertTrue(hasattr(image_processor, "image_mean"))
self.assertTrue(hasattr(image_processor, "image_std"))
self.assertTrue(hasattr(image_processor, "do_convert_rgb"))
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, numpify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, torchify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
@require_torch
@require_vision
class BlipImageProcessingTestFourChannels(FeatureExtractionSavingTestMixin, unittest.TestCase):
class BlipImageProcessingTestFourChannels(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = BlipImageProcessor if is_vision_available() else None
image_processing_class = BlipImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = BlipImageProcessingTester(self, num_channels=4)
self.image_processor_tester = BlipImageProcessingTester(self, num_channels=4)
self.expected_encoded_image_num_channels = 3
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_convert_rgb"))
def test_image_processor_properties(self):
image_processor = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processor, "do_resize"))
self.assertTrue(hasattr(image_processor, "size"))
self.assertTrue(hasattr(image_processor, "do_normalize"))
self.assertTrue(hasattr(image_processor, "image_mean"))
self.assertTrue(hasattr(image_processor, "image_std"))
self.assertTrue(hasattr(image_processor, "do_convert_rgb"))
def test_batch_feature(self):
pass
def test_call_pil_four_channels(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.image_processor_tester.batch_size,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)

174
tests/models/chinese_clip/test_image_processing_chinese_clip.py
View File

@@ -21,7 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import ImageProcessingSavingTestMixin
if is_torch_available():
@@ -30,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import ChineseCLIPFeatureExtractor
from transformers import ChineseCLIPImageProcessor
class ChineseCLIPFeatureExtractionTester(unittest.TestCase):
class ChineseCLIPImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -68,7 +68,7 @@ class ChineseCLIPFeatureExtractionTester(unittest.TestCase):
self.image_std = image_std
self.do_convert_rgb = do_convert_rgb
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -113,193 +113,193 @@ class ChineseCLIPFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class ChineseCLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class ChineseCLIPImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ChineseCLIPFeatureExtractor if is_vision_available() else None
image_processing_class = ChineseCLIPImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = ChineseCLIPFeatureExtractionTester(self, do_center_crop=True)
self.image_processor_tester = ChineseCLIPImageProcessingTester(self, do_center_crop=True)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_convert_rgb"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_convert_rgb"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 224, "width": 224})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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, {"height": 224, "width": 224})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, numpify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, torchify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
@require_torch
@require_vision
class ChineseCLIPFeatureExtractionTestFourChannels(FeatureExtractionSavingTestMixin, unittest.TestCase):
class ChineseCLIPImageProcessingTestFourChannels(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ChineseCLIPFeatureExtractor if is_vision_available() else None
image_processing_class = ChineseCLIPImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = ChineseCLIPFeatureExtractionTester(self, num_channels=4, do_center_crop=True)
self.image_processor_tester = ChineseCLIPImageProcessingTester(self, num_channels=4, do_center_crop=True)
self.expected_encoded_image_num_channels = 3
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_convert_rgb"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_convert_rgb"))
def test_batch_feature(self):
pass
def test_call_pil_four_channels(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.image_processor_tester.batch_size,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

174
tests/models/clip/test_image_processing_clip.py
View File

@@ -21,7 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import ImageProcessingSavingTestMixin
if is_torch_available():
@@ -30,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import CLIPFeatureExtractor
from transformers import CLIPImageProcessor
class CLIPFeatureExtractionTester(unittest.TestCase):
class CLIPImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -68,7 +68,7 @@ class CLIPFeatureExtractionTester(unittest.TestCase):
self.image_std = image_std
self.do_convert_rgb = do_convert_rgb
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -113,193 +113,193 @@ class CLIPFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class CLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class CLIPImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = CLIPFeatureExtractor if is_vision_available() else None
image_processing_class = CLIPImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = CLIPFeatureExtractionTester(self)
self.image_processor_tester = CLIPImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_convert_rgb"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_convert_rgb"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 20})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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": 20})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, numpify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, torchify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
@require_torch
@require_vision
class CLIPFeatureExtractionTestFourChannels(FeatureExtractionSavingTestMixin, unittest.TestCase):
class CLIPImageProcessingTestFourChannels(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = CLIPFeatureExtractor if is_vision_available() else None
image_processing_class = CLIPImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = CLIPFeatureExtractionTester(self, num_channels=4)
self.image_processor_tester = CLIPImageProcessingTester(self, num_channels=4)
self.expected_encoded_image_num_channels = 3
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_convert_rgb"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_convert_rgb"))
def test_batch_feature(self):
pass
def test_call_pil_four_channels(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.image_processor_tester.batch_size,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

133
tests/models/conditional_detr/test_image_processing_conditional_detr.py
View File

@@ -23,8 +23,7 @@ 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 ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -33,10 +32,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import ConditionalDetrFeatureExtractor
from transformers import ConditionalDetrImageProcessor
class ConditionalDetrFeatureExtractionTester(unittest.TestCase):
class ConditionalDetrImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -69,7 +68,7 @@ class ConditionalDetrFeatureExtractionTester(unittest.TestCase):
self.rescale_factor = rescale_factor
self.do_pad = do_pad
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -83,7 +82,7 @@ class ConditionalDetrFeatureExtractionTester(unittest.TestCase):
def get_expected_values(self, image_inputs, batched=False):
"""
This function computes the expected height and width when providing images to ConditionalDetrFeatureExtractor,
This function computes the expected height and width when providing images to ConditionalDetrImageProcessor,
assuming do_resize is set to True with a scalar size.
"""
if not batched:
@@ -115,149 +114,149 @@ class ConditionalDetrFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class ConditionalDetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class ConditionalDetrImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ConditionalDetrFeatureExtractor if is_vision_available() else None
image_processing_class = ConditionalDetrImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = ConditionalDetrFeatureExtractionTester(self)
self.image_processor_tester = ConditionalDetrImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
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_resize"))
self.assertTrue(hasattr(image_processing, "size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 18, "longest_edge": 1333})
self.assertEqual(feature_extractor.do_pad, True)
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)
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, max_size=84, pad_and_return_pixel_mask=False
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(feature_extractor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(feature_extractor.do_pad, False)
self.assertEqual(image_processor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(image_processor.do_pad, False)
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_equivalence_pad_and_create_pixel_mask(self):
# Initialize feature_extractors
feature_extractor_1 = self.feature_extraction_class(**self.feat_extract_dict)
feature_extractor_2 = self.feature_extraction_class(do_resize=False, do_normalize=False, do_rescale=False)
# Initialize image_processings
image_processing_1 = self.image_processing_class(**self.image_processor_dict)
image_processing_2 = self.image_processing_class(do_resize=False, do_normalize=False, do_rescale=False)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test whether the method "pad_and_return_pixel_mask" and calling the feature extractor return the same tensors
encoded_images_with_method = feature_extractor_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = feature_extractor_2(image_inputs, return_tensors="pt")
# Test whether the method "pad_and_return_pixel_mask" and calling the image processor return the same tensors
encoded_images_with_method = image_processing_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = image_processing_2(image_inputs, return_tensors="pt")
self.assertTrue(
torch.allclose(encoded_images_with_method["pixel_values"], encoded_images["pixel_values"], atol=1e-4)
@@ -276,8 +275,8 @@ class ConditionalDetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, uni
target = {"image_id": 39769, "annotations": target}
# encode them
feature_extractor = ConditionalDetrFeatureExtractor.from_pretrained("microsoft/conditional-detr-resnet-50")
encoding = feature_extractor(images=image, annotations=target, return_tensors="pt")
image_processing = ConditionalDetrImageProcessor.from_pretrained("microsoft/conditional-detr-resnet-50")
encoding = image_processing(images=image, annotations=target, return_tensors="pt")
# verify pixel values
expected_shape = torch.Size([1, 3, 800, 1066])
@@ -322,8 +321,8 @@ class ConditionalDetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, uni
masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
# encode them
feature_extractor = ConditionalDetrFeatureExtractor(format="coco_panoptic")
encoding = feature_extractor(images=image, annotations=target, masks_path=masks_path, return_tensors="pt")
image_processing = ConditionalDetrImageProcessor(format="coco_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])

117
tests/models/convnext/test_image_processing_convnext.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import ConvNextFeatureExtractor
from transformers import ConvNextImageProcessor
class ConvNextFeatureExtractionTester(unittest.TestCase):
class ConvNextImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -64,7 +63,7 @@ class ConvNextFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -77,128 +76,128 @@ class ConvNextFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class ConvNextFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class ConvNextImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ConvNextFeatureExtractor if is_vision_available() else None
image_processing_class = ConvNextImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = ConvNextFeatureExtractionTester(self)
self.image_processor_tester = ConvNextImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "crop_pct"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "crop_pct"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 20})
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": 20})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["shortest_edge"],
self.image_processor_tester.size["shortest_edge"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["shortest_edge"],
self.image_processor_tester.size["shortest_edge"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["shortest_edge"],
self.image_processor_tester.size["shortest_edge"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["shortest_edge"],
self.image_processor_tester.size["shortest_edge"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["shortest_edge"],
self.image_processor_tester.size["shortest_edge"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["shortest_edge"],
self.image_processor_tester.size["shortest_edge"],
),
)

137
tests/models/deformable_detr/test_image_processing_deformable_detr.py
View File

@@ -23,8 +23,7 @@ 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 ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -33,10 +32,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import DeformableDetrFeatureExtractor
from transformers import DeformableDetrImageProcessor
class DeformableDetrFeatureExtractionTester(unittest.TestCase):
class DeformableDetrImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -69,7 +68,7 @@ class DeformableDetrFeatureExtractionTester(unittest.TestCase):
self.rescale_factor = rescale_factor
self.do_pad = do_pad
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -83,7 +82,7 @@ class DeformableDetrFeatureExtractionTester(unittest.TestCase):
def get_expected_values(self, image_inputs, batched=False):
"""
This function computes the expected height and width when providing images to DeformableDetrFeatureExtractor,
This function computes the expected height and width when providing images to DeformableDetrImageProcessor,
assuming do_resize is set to True with a scalar size.
"""
if not batched:
@@ -115,152 +114,152 @@ class DeformableDetrFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class DeformableDetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class DeformableDetrImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = DeformableDetrFeatureExtractor if is_vision_available() else None
image_processing_class = DeformableDetrImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = DeformableDetrFeatureExtractionTester(self)
self.image_processor_tester = DeformableDetrImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "do_rescale"))
self.assertTrue(hasattr(feature_extractor, "do_pad"))
self.assertTrue(hasattr(feature_extractor, "size"))
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_resize"))
self.assertTrue(hasattr(image_processing, "do_rescale"))
self.assertTrue(hasattr(image_processing, "do_pad"))
self.assertTrue(hasattr(image_processing, "size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 18, "longest_edge": 1333})
self.assertEqual(feature_extractor.do_pad, True)
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)
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, max_size=84, pad_and_return_pixel_mask=False
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(feature_extractor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(feature_extractor.do_pad, False)
self.assertEqual(image_processor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(image_processor.do_pad, False)
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_equivalence_pad_and_create_pixel_mask(self):
# Initialize feature_extractors
feature_extractor_1 = self.feature_extraction_class(**self.feat_extract_dict)
feature_extractor_2 = self.feature_extraction_class(do_resize=False, do_normalize=False, do_rescale=False)
# Initialize image_processings
image_processing_1 = self.image_processing_class(**self.image_processor_dict)
image_processing_2 = self.image_processing_class(do_resize=False, do_normalize=False, do_rescale=False)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test whether the method "pad_and_return_pixel_mask" and calling the feature extractor return the same tensors
encoded_images_with_method = feature_extractor_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = feature_extractor_2(image_inputs, return_tensors="pt")
# Test whether the method "pad_and_return_pixel_mask" and calling the image processor return the same tensors
encoded_images_with_method = image_processing_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = image_processing_2(image_inputs, return_tensors="pt")
self.assertTrue(
torch.allclose(encoded_images_with_method["pixel_values"], encoded_images["pixel_values"], atol=1e-4)
@@ -279,8 +278,8 @@ class DeformableDetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, unit
target = {"image_id": 39769, "annotations": target}
# encode them
feature_extractor = DeformableDetrFeatureExtractor()
encoding = feature_extractor(images=image, annotations=target, return_tensors="pt")
image_processing = DeformableDetrImageProcessor()
encoding = image_processing(images=image, annotations=target, return_tensors="pt")
# verify pixel values
expected_shape = torch.Size([1, 3, 800, 1066])
@@ -325,8 +324,8 @@ class DeformableDetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, unit
masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
# encode them
feature_extractor = DeformableDetrFeatureExtractor(format="coco_panoptic")
encoding = feature_extractor(images=image, annotations=target, masks_path=masks_path, return_tensors="pt")
image_processing = DeformableDetrImageProcessor(format="coco_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])

123
tests/models/deit/test_image_processing_deit.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import DeiTFeatureExtractor
from transformers import DeiTImageProcessor
class DeiTFeatureExtractionTester(unittest.TestCase):
class DeiTImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -68,7 +67,7 @@ class DeiTFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -82,132 +81,132 @@ class DeiTFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class DeiTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class DeiTImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = DeiTFeatureExtractor if is_vision_available() else None
image_processing_class = DeiTImageProcessor if is_vision_available() else None
test_cast_dtype = True
def setUp(self):
self.feature_extract_tester = DeiTFeatureExtractionTester(self)
self.image_processor_tester = DeiTImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 20, "width": 20})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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, {"height": 20, "width": 20})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

139
tests/models/detr/test_image_processing_detr.py
View File

@@ -23,8 +23,7 @@ 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 ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -33,10 +32,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import DetrFeatureExtractor
from transformers import DetrImageProcessor
class DetrFeatureExtractionTester(unittest.TestCase):
class DetrImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -69,7 +68,7 @@ class DetrFeatureExtractionTester(unittest.TestCase):
self.image_std = image_std
self.do_pad = do_pad
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -83,7 +82,7 @@ class DetrFeatureExtractionTester(unittest.TestCase):
def get_expected_values(self, image_inputs, batched=False):
"""
This function computes the expected height and width when providing images to DetrFeatureExtractor,
This function computes the expected height and width when providing images to DetrImageProcessor,
assuming do_resize is set to True with a scalar size.
"""
if not batched:
@@ -115,152 +114,152 @@ class DetrFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class DetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class DetrImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = DetrFeatureExtractor if is_vision_available() else None
image_processing_class = DetrImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = DetrFeatureExtractionTester(self)
self.image_processor_tester = DetrImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_rescale"))
self.assertTrue(hasattr(feature_extractor, "rescale_factor"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_pad"))
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"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 18, "longest_edge": 1333})
self.assertEqual(feature_extractor.do_pad, True)
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)
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, max_size=84, pad_and_return_pixel_mask=False
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(feature_extractor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(feature_extractor.do_pad, False)
self.assertEqual(image_processor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(image_processor.do_pad, False)
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_equivalence_pad_and_create_pixel_mask(self):
# Initialize feature_extractors
feature_extractor_1 = self.feature_extraction_class(**self.feat_extract_dict)
feature_extractor_2 = self.feature_extraction_class(do_resize=False, do_normalize=False, do_rescale=False)
# Initialize image_processings
image_processing_1 = self.image_processing_class(**self.image_processor_dict)
image_processing_2 = self.image_processing_class(do_resize=False, do_normalize=False, do_rescale=False)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test whether the method "pad_and_return_pixel_mask" and calling the feature extractor return the same tensors
encoded_images_with_method = feature_extractor_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = feature_extractor_2(image_inputs, return_tensors="pt")
# Test whether the method "pad_and_return_pixel_mask" and calling the image processor return the same tensors
encoded_images_with_method = image_processing_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = image_processing_2(image_inputs, return_tensors="pt")
self.assertTrue(
torch.allclose(encoded_images_with_method["pixel_values"], encoded_images["pixel_values"], atol=1e-4)
@@ -279,8 +278,8 @@ class DetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
target = {"image_id": 39769, "annotations": target}
# encode them
feature_extractor = DetrFeatureExtractor.from_pretrained("facebook/detr-resnet-50")
encoding = feature_extractor(images=image, annotations=target, return_tensors="pt")
image_processing = DetrImageProcessor.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])
@@ -325,8 +324,8 @@ class DetrFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
# encode them
feature_extractor = DetrFeatureExtractor.from_pretrained("facebook/detr-resnet-50-panoptic")
encoding = feature_extractor(images=image, annotations=target, masks_path=masks_path, return_tensors="pt")
image_processing = DetrImageProcessor.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])

125
tests/models/donut/test_image_processing_donut.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import is_flaky, require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import DonutFeatureExtractor
from transformers import DonutImageProcessor
class DonutFeatureExtractionTester(unittest.TestCase):
class DonutImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -67,7 +66,7 @@ class DonutFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -82,137 +81,137 @@ class DonutFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class DonutFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class DonutImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = DonutFeatureExtractor if is_vision_available() else None
image_processing_class = DonutImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = DonutFeatureExtractionTester(self)
self.image_processor_tester = DonutImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_thumbnail"))
self.assertTrue(hasattr(feature_extractor, "do_align_long_axis"))
self.assertTrue(hasattr(feature_extractor, "do_pad"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_thumbnail"))
self.assertTrue(hasattr(image_processing, "do_align_long_axis"))
self.assertTrue(hasattr(image_processing, "do_pad"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 18, "width": 20})
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, {"height": 18, "width": 20})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
# Previous config had dimensions in (width, height) order
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=(42, 84))
self.assertEqual(feature_extractor.size, {"height": 84, "width": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=(42, 84))
self.assertEqual(image_processor.size, {"height": 84, "width": 42})
def test_batch_feature(self):
pass
@is_flaky()
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
@is_flaky()
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
@is_flaky()
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)

115
tests/models/dpt/test_image_processing_dpt.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.file_utils import is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_vision
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import DPTFeatureExtractor
from transformers import DPTImageProcessor
class DPTFeatureExtractionTester(unittest.TestCase):
class DPTImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -62,7 +61,7 @@ class DPTFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -74,124 +73,124 @@ class DPTFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class DPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class DPTImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = DPTFeatureExtractor if is_vision_available() else None
image_processing_class = DPTImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = DPTFeatureExtractionTester(self)
self.image_processor_tester = DPTImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
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_resize"))
self.assertTrue(hasattr(image_processing, "size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 18, "width": 18})
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, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)

103
tests/models/efficientformer/test_image_processing_efficientformer.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,7 +30,7 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import ViTFeatureExtractor
from transformers import ViTImageProcessor
class EfficientFormerImageProcessorTester(unittest.TestCase):
@@ -62,7 +61,7 @@ class EfficientFormerImageProcessorTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -74,120 +73,120 @@ class EfficientFormerImageProcessorTester(unittest.TestCase):
@require_torch
@require_vision
class EfficientFormerImageProcessorTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class EfficientFormerImageProcessorTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ViTFeatureExtractor if is_vision_available() else None
image_processing_class = ViTImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = EfficientFormerImageProcessorTester(self)
self.image_proc_tester = EfficientFormerImageProcessorTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_proc_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
def test_image_proc_properties(self):
image_processor = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processor, "image_mean"))
self.assertTrue(hasattr(image_processor, "image_std"))
self.assertTrue(hasattr(image_processor, "do_normalize"))
self.assertTrue(hasattr(image_processor, "do_resize"))
self.assertTrue(hasattr(image_processor, "size"))
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_proc_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_proc_tester.num_channels,
self.image_proc_tester.size["height"],
self.image_proc_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_proc_tester.batch_size,
self.image_proc_tester.num_channels,
self.image_proc_tester.size["height"],
self.image_proc_tester.size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_proc_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_proc_tester.num_channels,
self.image_proc_tester.size["height"],
self.image_proc_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_proc_tester.batch_size,
self.image_proc_tester.num_channels,
self.image_proc_tester.size["height"],
self.image_proc_tester.size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_proc_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_proc_tester.num_channels,
self.image_proc_tester.size["height"],
self.image_proc_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_proc_tester.batch_size,
self.image_proc_tester.num_channels,
self.image_proc_tester.size["height"],
self.image_proc_tester.size["width"],
),
)

177
tests/models/flava/test_image_processing_flava.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,7 +30,7 @@ if is_torch_available():
if is_vision_available():
import PIL
from transformers import FlavaFeatureExtractor
from transformers import FlavaImageProcessor
from transformers.image_utils import PILImageResampling
from transformers.models.flava.image_processing_flava import (
FLAVA_CODEBOOK_MEAN,
@@ -43,7 +42,7 @@ else:
FLAVA_IMAGE_MEAN = FLAVA_IMAGE_STD = FLAVA_CODEBOOK_MEAN = FLAVA_CODEBOOK_STD = None
class FlavaFeatureExtractionTester(unittest.TestCase):
class FlavaImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -115,7 +114,7 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
self.codebook_image_mean = codebook_image_mean
self.codebook_image_std = codebook_image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -160,82 +159,82 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class FlavaFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class FlavaImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = FlavaFeatureExtractor if is_vision_available() else None
image_processing_class = FlavaImageProcessor if is_vision_available() else None
maxDiff = None
def setUp(self):
self.feature_extract_tester = FlavaFeatureExtractionTester(self)
self.image_processor_tester = FlavaImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "resample"))
self.assertTrue(hasattr(feature_extractor, "crop_size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_rescale"))
self.assertTrue(hasattr(feature_extractor, "rescale_factor"))
self.assertTrue(hasattr(feature_extractor, "masking_generator"))
self.assertTrue(hasattr(feature_extractor, "codebook_do_resize"))
self.assertTrue(hasattr(feature_extractor, "codebook_size"))
self.assertTrue(hasattr(feature_extractor, "codebook_resample"))
self.assertTrue(hasattr(feature_extractor, "codebook_do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "codebook_crop_size"))
self.assertTrue(hasattr(feature_extractor, "codebook_do_map_pixels"))
self.assertTrue(hasattr(feature_extractor, "codebook_do_normalize"))
self.assertTrue(hasattr(feature_extractor, "codebook_image_mean"))
self.assertTrue(hasattr(feature_extractor, "codebook_image_std"))
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_resize"))
self.assertTrue(hasattr(image_processing, "resample"))
self.assertTrue(hasattr(image_processing, "crop_size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "do_rescale"))
self.assertTrue(hasattr(image_processing, "rescale_factor"))
self.assertTrue(hasattr(image_processing, "masking_generator"))
self.assertTrue(hasattr(image_processing, "codebook_do_resize"))
self.assertTrue(hasattr(image_processing, "codebook_size"))
self.assertTrue(hasattr(image_processing, "codebook_resample"))
self.assertTrue(hasattr(image_processing, "codebook_do_center_crop"))
self.assertTrue(hasattr(image_processing, "codebook_crop_size"))
self.assertTrue(hasattr(image_processing, "codebook_do_map_pixels"))
self.assertTrue(hasattr(image_processing, "codebook_do_normalize"))
self.assertTrue(hasattr(image_processing, "codebook_image_mean"))
self.assertTrue(hasattr(image_processing, "codebook_image_std"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 224, "width": 224})
self.assertEqual(feature_extractor.crop_size, {"height": 224, "width": 224})
self.assertEqual(feature_extractor.codebook_size, {"height": 112, "width": 112})
self.assertEqual(feature_extractor.codebook_crop_size, {"height": 112, "width": 112})
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, {"height": 224, "width": 224})
self.assertEqual(image_processor.crop_size, {"height": 224, "width": 224})
self.assertEqual(image_processor.codebook_size, {"height": 112, "width": 112})
self.assertEqual(image_processor.codebook_crop_size, {"height": 112, "width": 112})
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, crop_size=84, codebook_size=33, codebook_crop_size=66
image_processor = self.image_processing_class.from_dict(
self.image_processor_dict, size=42, crop_size=84, codebook_size=33, codebook_crop_size=66
)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
self.assertEqual(feature_extractor.codebook_size, {"height": 33, "width": 33})
self.assertEqual(feature_extractor.codebook_crop_size, {"height": 66, "width": 66})
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
self.assertEqual(image_processor.codebook_size, {"height": 33, "width": 33})
self.assertEqual(image_processor.codebook_crop_size, {"height": 66, "width": 66})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, PIL.Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt")
encoded_images = image_processing(image_inputs[0], return_tensors="pt")
# Test no bool masked pos
self.assertFalse("bool_masked_pos" in encoded_images)
expected_height, expected_width = self.feature_extract_tester.get_expected_image_size()
expected_height, expected_width = self.image_processor_tester.get_expected_image_size()
self.assertEqual(
encoded_images.pixel_values.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt")
expected_height, expected_width = self.feature_extract_tester.get_expected_image_size()
encoded_images = image_processing(image_inputs, return_tensors="pt")
expected_height, expected_width = self.image_processor_tester.get_expected_image_size()
# Test no bool masked pos
self.assertFalse("bool_masked_pos" in encoded_images)
@@ -243,86 +242,86 @@ class FlavaFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
self.assertEqual(
encoded_images.pixel_values.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def _test_call_framework(self, instance_class, prepare_kwargs):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, **prepare_kwargs)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, **prepare_kwargs)
for image in image_inputs:
self.assertIsInstance(image, instance_class)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt")
encoded_images = image_processing(image_inputs[0], return_tensors="pt")
expected_height, expected_width = self.feature_extract_tester.get_expected_image_size()
expected_height, expected_width = self.image_processor_tester.get_expected_image_size()
self.assertEqual(
encoded_images.pixel_values.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
encoded_images = feature_extractor(image_inputs, return_image_mask=True, return_tensors="pt")
encoded_images = image_processing(image_inputs, return_image_mask=True, return_tensors="pt")
expected_height, expected_width = self.feature_extract_tester.get_expected_image_size()
expected_height, expected_width = self.image_processor_tester.get_expected_image_size()
self.assertEqual(
encoded_images.pixel_values.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
expected_height, expected_width = self.feature_extract_tester.get_expected_mask_size()
expected_height, expected_width = self.image_processor_tester.get_expected_mask_size()
self.assertEqual(
encoded_images.bool_masked_pos.shape,
(
self.feature_extract_tester.batch_size,
self.image_processor_tester.batch_size,
expected_height,
expected_width,
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_image_size()
expected_height, expected_width = self.image_processor_tester.get_expected_image_size()
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
# Test masking
encoded_images = feature_extractor(image_inputs, return_image_mask=True, return_tensors="pt")
encoded_images = image_processing(image_inputs, return_image_mask=True, return_tensors="pt")
expected_height, expected_width = self.feature_extract_tester.get_expected_image_size()
expected_height, expected_width = self.image_processor_tester.get_expected_image_size()
self.assertEqual(
encoded_images.pixel_values.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
expected_height, expected_width = self.feature_extract_tester.get_expected_mask_size()
expected_height, expected_width = self.image_processor_tester.get_expected_mask_size()
self.assertEqual(
encoded_images.bool_masked_pos.shape,
(
self.feature_extract_tester.batch_size,
self.image_processor_tester.batch_size,
expected_height,
expected_width,
),
@@ -335,39 +334,39 @@ class FlavaFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
self._test_call_framework(torch.Tensor, prepare_kwargs={"torchify": True})
def test_masking(self):
# Initialize feature_extractor
# Initialize image_processing
random.seed(1234)
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_processing = self.image_processing_class(**self.image_processor_dict)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_image_mask=True, return_tensors="pt")
encoded_images = image_processing(image_inputs[0], return_image_mask=True, return_tensors="pt")
self.assertEqual(encoded_images.bool_masked_pos.sum().item(), 75)
def test_codebook_pixels(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, PIL.Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_codebook_pixels=True, return_tensors="pt")
expected_height, expected_width = self.feature_extract_tester.get_expected_codebook_image_size()
encoded_images = image_processing(image_inputs[0], return_codebook_pixels=True, return_tensors="pt")
expected_height, expected_width = self.image_processor_tester.get_expected_codebook_image_size()
self.assertEqual(
encoded_images.codebook_pixel_values.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_codebook_pixels=True, return_tensors="pt")
expected_height, expected_width = self.feature_extract_tester.get_expected_codebook_image_size()
encoded_images = image_processing(image_inputs, return_codebook_pixels=True, return_tensors="pt")
expected_height, expected_width = self.image_processor_tester.get_expected_codebook_image_size()
self.assertEqual(
encoded_images.codebook_pixel_values.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),

73
tests/models/glpn/test_image_processing_glpn.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import GLPNFeatureExtractor
from transformers import GLPNImageProcessor
class GLPNFeatureExtractionTester(unittest.TestCase):
class GLPNImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -57,7 +56,7 @@ class GLPNFeatureExtractionTester(unittest.TestCase):
self.size_divisor = size_divisor
self.do_rescale = do_rescale
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size_divisor": self.size_divisor,
@@ -67,62 +66,62 @@ class GLPNFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class GLPNFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class GLPNImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = GLPNFeatureExtractor if is_vision_available() else None
image_processing_class = GLPNImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = GLPNFeatureExtractionTester(self)
self.image_processor_tester = GLPNImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size_divisor"))
self.assertTrue(hasattr(feature_extractor, "resample"))
self.assertTrue(hasattr(feature_extractor, "do_rescale"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size_divisor"))
self.assertTrue(hasattr(image_processing, "resample"))
self.assertTrue(hasattr(image_processing, "do_rescale"))
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input (GLPNFeatureExtractor doesn't support batching)
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
self.assertTrue(encoded_images.shape[-1] % self.feature_extract_tester.size_divisor == 0)
self.assertTrue(encoded_images.shape[-2] % self.feature_extract_tester.size_divisor == 0)
# Test not batched input (GLPNImageProcessor doesn't support batching)
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertTrue(encoded_images.shape[-1] % self.image_processor_tester.size_divisor == 0)
self.assertTrue(encoded_images.shape[-2] % self.image_processor_tester.size_divisor == 0)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input (GLPNFeatureExtractor doesn't support batching)
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
self.assertTrue(encoded_images.shape[-1] % self.feature_extract_tester.size_divisor == 0)
self.assertTrue(encoded_images.shape[-2] % self.feature_extract_tester.size_divisor == 0)
# Test not batched input (GLPNImageProcessor doesn't support batching)
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertTrue(encoded_images.shape[-1] % self.image_processor_tester.size_divisor == 0)
self.assertTrue(encoded_images.shape[-2] % self.image_processor_tester.size_divisor == 0)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input (GLPNFeatureExtractor doesn't support batching)
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
self.assertTrue(encoded_images.shape[-1] % self.feature_extract_tester.size_divisor == 0)
self.assertTrue(encoded_images.shape[-2] % self.feature_extract_tester.size_divisor == 0)
# Test not batched input (GLPNImageProcessor doesn't support batching)
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertTrue(encoded_images.shape[-1] % self.image_processor_tester.size_divisor == 0)
self.assertTrue(encoded_images.shape[-2] % self.image_processor_tester.size_divisor == 0)

90
tests/models/imagegpt/test_image_processing_imagegpt.py
View File

@@ -25,7 +25,7 @@ from datasets import load_dataset
from transformers.testing_utils import require_torch, require_vision, slow
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import ImageProcessingSavingTestMixin
if is_torch_available():
@@ -34,10 +34,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import ImageGPTFeatureExtractor
from transformers import ImageGPTImageProcessor
class ImageGPTFeatureExtractionTester(unittest.TestCase):
class ImageGPTImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -61,7 +61,7 @@ class ImageGPTFeatureExtractionTester(unittest.TestCase):
self.size = size
self.do_normalize = do_normalize
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
# here we create 2 clusters for the sake of simplicity
"clusters": np.asarray(
@@ -78,68 +78,68 @@ class ImageGPTFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class ImageGPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class ImageGPTImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ImageGPTFeatureExtractor if is_vision_available() else None
image_processing_class = ImageGPTImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = ImageGPTFeatureExtractionTester(self)
self.image_processor_tester = ImageGPTImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "clusters"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "clusters"))
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 18, "width": 18})
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, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
def test_feat_extract_to_json_string(self):
feat_extract = self.feature_extraction_class(**self.feat_extract_dict)
obj = json.loads(feat_extract.to_json_string())
for key, value in self.feat_extract_dict.items():
def test_image_processor_to_json_string(self):
image_processor = self.image_processing_class(**self.image_processor_dict)
obj = json.loads(image_processor.to_json_string())
for key, value in self.image_processor_dict.items():
if key == "clusters":
self.assertTrue(np.array_equal(value, obj[key]))
else:
self.assertEqual(obj[key], value)
def test_feat_extract_to_json_file(self):
feat_extract_first = self.feature_extraction_class(**self.feat_extract_dict)
def test_image_processor_to_json_file(self):
image_processor_first = self.image_processing_class(**self.image_processor_dict)
with tempfile.TemporaryDirectory() as tmpdirname:
json_file_path = os.path.join(tmpdirname, "feat_extract.json")
feat_extract_first.to_json_file(json_file_path)
feat_extract_second = self.feature_extraction_class.from_json_file(json_file_path).to_dict()
json_file_path = os.path.join(tmpdirname, "image_processor.json")
image_processor_first.to_json_file(json_file_path)
image_processor_second = self.image_processing_class.from_json_file(json_file_path).to_dict()
feat_extract_first = feat_extract_first.to_dict()
for key, value in feat_extract_first.items():
image_processor_first = image_processor_first.to_dict()
for key, value in image_processor_first.items():
if key == "clusters":
self.assertTrue(np.array_equal(value, feat_extract_second[key]))
self.assertTrue(np.array_equal(value, image_processor_second[key]))
else:
self.assertEqual(feat_extract_first[key], value)
self.assertEqual(image_processor_first[key], value)
def test_feat_extract_from_and_save_pretrained(self):
feat_extract_first = self.feature_extraction_class(**self.feat_extract_dict)
def test_image_processor_from_and_save_pretrained(self):
image_processor_first = self.image_processing_class(**self.image_processor_dict)
with tempfile.TemporaryDirectory() as tmpdirname:
feat_extract_first.save_pretrained(tmpdirname)
feat_extract_second = self.feature_extraction_class.from_pretrained(tmpdirname).to_dict()
image_processor_first.save_pretrained(tmpdirname)
image_processor_second = self.image_processing_class.from_pretrained(tmpdirname).to_dict()
feat_extract_first = feat_extract_first.to_dict()
for key, value in feat_extract_first.items():
image_processor_first = image_processor_first.to_dict()
for key, value in image_processor_first.items():
if key == "clusters":
self.assertTrue(np.array_equal(value, feat_extract_second[key]))
self.assertTrue(np.array_equal(value, image_processor_second[key]))
else:
self.assertEqual(feat_extract_first[key], value)
self.assertEqual(image_processor_first[key], value)
@unittest.skip("ImageGPT requires clusters at initialization")
def test_init_without_params(self):
@@ -159,15 +159,15 @@ def prepare_images():
@require_vision
@require_torch
class ImageGPTFeatureExtractorIntegrationTest(unittest.TestCase):
class ImageGPTImageProcessorIntegrationTest(unittest.TestCase):
@slow
def test_image(self):
feature_extractor = ImageGPTFeatureExtractor.from_pretrained("openai/imagegpt-small")
image_processing = ImageGPTImageProcessor.from_pretrained("openai/imagegpt-small")
images = prepare_images()
# test non-batched
encoding = feature_extractor(images[0], return_tensors="pt")
encoding = image_processing(images[0], return_tensors="pt")
self.assertIsInstance(encoding.input_ids, torch.LongTensor)
self.assertEqual(encoding.input_ids.shape, (1, 1024))
@@ -176,7 +176,7 @@ class ImageGPTFeatureExtractorIntegrationTest(unittest.TestCase):
self.assertEqual(encoding.input_ids[0, :3].tolist(), expected_slice)
# test batched
encoding = feature_extractor(images, return_tensors="pt")
encoding = image_processing(images, return_tensors="pt")
self.assertIsInstance(encoding.input_ids, torch.LongTensor)
self.assertEqual(encoding.input_ids.shape, (2, 1024))

119
tests/models/layoutlmv2/test_image_processing_layoutlmv2.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_pytesseract, require_torch
from transformers.utils import is_pytesseract_available, is_torch_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_pytesseract_available():
from PIL import Image
from transformers import LayoutLMv2FeatureExtractor
from transformers import LayoutLMv2ImageProcessor
class LayoutLMv2FeatureExtractionTester(unittest.TestCase):
class LayoutLMv2ImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -58,56 +57,56 @@ class LayoutLMv2FeatureExtractionTester(unittest.TestCase):
self.size = size
self.apply_ocr = apply_ocr
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {"do_resize": self.do_resize, "size": self.size, "apply_ocr": self.apply_ocr}
@require_torch
@require_pytesseract
class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class LayoutLMv2ImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = LayoutLMv2FeatureExtractor if is_pytesseract_available() else None
image_processing_class = LayoutLMv2ImageProcessor if is_pytesseract_available() else None
def setUp(self):
self.feature_extract_tester = LayoutLMv2FeatureExtractionTester(self)
self.image_processor_tester = LayoutLMv2ImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "apply_ocr"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "apply_ocr"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 18, "width": 18})
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, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoding = feature_extractor(image_inputs[0], return_tensors="pt")
encoding = image_processing(image_inputs[0], return_tensors="pt")
self.assertEqual(
encoding.pixel_values.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
@@ -115,84 +114,84 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
self.assertIsInstance(encoding.boxes, list)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_layoutlmv2_integration_test(self):
# with apply_OCR = True
feature_extractor = LayoutLMv2FeatureExtractor()
image_processing = LayoutLMv2ImageProcessor()
from datasets import load_dataset
@@ -200,7 +199,7 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
image = Image.open(ds[0]["file"]).convert("RGB")
encoding = feature_extractor(image, return_tensors="pt")
encoding = image_processing(image, return_tensors="pt")
self.assertEqual(encoding.pixel_values.shape, (1, 3, 224, 224))
self.assertEqual(len(encoding.words), len(encoding.boxes))
@@ -215,8 +214,8 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
self.assertListEqual(encoding.boxes, expected_boxes)
# with apply_OCR = False
feature_extractor = LayoutLMv2FeatureExtractor(apply_ocr=False)
image_processing = LayoutLMv2ImageProcessor(apply_ocr=False)
encoding = feature_extractor(image, return_tensors="pt")
encoding = image_processing(image, return_tensors="pt")
self.assertEqual(encoding.pixel_values.shape, (1, 3, 224, 224))

119
tests/models/layoutlmv3/test_image_processing_layoutlmv3.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_pytesseract, require_torch
from transformers.utils import is_pytesseract_available, is_torch_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_pytesseract_available():
from PIL import Image
from transformers import LayoutLMv3FeatureExtractor
from transformers import LayoutLMv3ImageProcessor
class LayoutLMv3FeatureExtractionTester(unittest.TestCase):
class LayoutLMv3ImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -58,56 +57,56 @@ class LayoutLMv3FeatureExtractionTester(unittest.TestCase):
self.size = size
self.apply_ocr = apply_ocr
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {"do_resize": self.do_resize, "size": self.size, "apply_ocr": self.apply_ocr}
@require_torch
@require_pytesseract
class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class LayoutLMv3ImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = LayoutLMv3FeatureExtractor if is_pytesseract_available() else None
image_processing_class = LayoutLMv3ImageProcessor if is_pytesseract_available() else None
def setUp(self):
self.feature_extract_tester = LayoutLMv3FeatureExtractionTester(self)
self.image_processor_tester = LayoutLMv3ImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "apply_ocr"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "apply_ocr"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 18, "width": 18})
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, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoding = feature_extractor(image_inputs[0], return_tensors="pt")
encoding = image_processing(image_inputs[0], return_tensors="pt")
self.assertEqual(
encoding.pixel_values.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
@@ -115,84 +114,84 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
self.assertIsInstance(encoding.boxes, list)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_LayoutLMv3_integration_test(self):
# with apply_OCR = True
feature_extractor = LayoutLMv3FeatureExtractor()
image_processing = LayoutLMv3ImageProcessor()
from datasets import load_dataset
@@ -200,7 +199,7 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
image = Image.open(ds[0]["file"]).convert("RGB")
encoding = feature_extractor(image, return_tensors="pt")
encoding = image_processing(image, return_tensors="pt")
self.assertEqual(encoding.pixel_values.shape, (1, 3, 224, 224))
self.assertEqual(len(encoding.words), len(encoding.boxes))
@@ -215,8 +214,8 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
self.assertListEqual(encoding.boxes, expected_boxes)
# with apply_OCR = False
feature_extractor = LayoutLMv3FeatureExtractor(apply_ocr=False)
image_processing = LayoutLMv3ImageProcessor(apply_ocr=False)
encoding = feature_extractor(image, return_tensors="pt")
encoding = image_processing(image, return_tensors="pt")
self.assertEqual(encoding.pixel_values.shape, (1, 3, 224, 224))

121
tests/models/levit/test_image_processing_levit.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import LevitFeatureExtractor
from transformers import LevitImageProcessor
class LevitFeatureExtractionTester(unittest.TestCase):
class LevitImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -67,7 +66,7 @@ class LevitFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -81,130 +80,130 @@ class LevitFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class LevitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class LevitImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = LevitFeatureExtractor if is_vision_available() else None
image_processing_class = LevitImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = LevitFeatureExtractionTester(self)
self.image_processor_tester = LevitImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "size"))
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_resize"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 18})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

213
tests/models/maskformer/test_image_processing_maskformer.py
View File

@@ -23,15 +23,14 @@ from huggingface_hub import hf_hub_download
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from transformers import MaskFormerFeatureExtractor
from transformers import MaskFormerImageProcessor
from transformers.models.maskformer.image_processing_maskformer import binary_mask_to_rle
from transformers.models.maskformer.modeling_maskformer import MaskFormerForInstanceSegmentationOutput
@@ -39,7 +38,7 @@ if is_vision_available():
from PIL import Image
class MaskFormerFeatureExtractionTester(unittest.TestCase):
class MaskFormerImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -77,7 +76,7 @@ class MaskFormerFeatureExtractionTester(unittest.TestCase):
self.do_reduce_labels = do_reduce_labels
self.ignore_index = ignore_index
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -92,7 +91,7 @@ class MaskFormerFeatureExtractionTester(unittest.TestCase):
def get_expected_values(self, image_inputs, batched=False):
"""
This function computes the expected height and width when providing images to MaskFormerFeatureExtractor,
This function computes the expected height and width when providing images to MaskFormerImageProcessor,
assuming do_resize is set to True with a scalar size.
"""
if not batched:
@@ -131,154 +130,154 @@ class MaskFormerFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class MaskFormerImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = MaskFormerFeatureExtractor if (is_vision_available() and is_torch_available()) else None
image_processing_class = MaskFormerImageProcessor if (is_vision_available() and is_torch_available()) else None
def setUp(self):
self.feature_extract_tester = MaskFormerFeatureExtractionTester(self)
self.image_processor_tester = MaskFormerImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "max_size"))
self.assertTrue(hasattr(feature_extractor, "ignore_index"))
self.assertTrue(hasattr(feature_extractor, "num_labels"))
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_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "max_size"))
self.assertTrue(hasattr(image_processing, "ignore_index"))
self.assertTrue(hasattr(image_processing, "num_labels"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 32, "longest_edge": 1333})
self.assertEqual(feature_extractor.size_divisor, 0)
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": 32, "longest_edge": 1333})
self.assertEqual(image_processor.size_divisor, 0)
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, max_size=84, size_divisibility=8
image_processor = self.image_processing_class.from_dict(
self.image_processor_dict, size=42, max_size=84, size_divisibility=8
)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(feature_extractor.size_divisor, 8)
self.assertEqual(image_processor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(image_processor.size_divisor, 8)
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_equivalence_pad_and_create_pixel_mask(self):
# Initialize feature_extractors
feature_extractor_1 = self.feature_extraction_class(**self.feat_extract_dict)
feature_extractor_2 = self.feature_extraction_class(
do_resize=False, do_normalize=False, do_rescale=False, num_labels=self.feature_extract_tester.num_classes
# Initialize image_processings
image_processing_1 = self.image_processing_class(**self.image_processor_dict)
image_processing_2 = self.image_processing_class(
do_resize=False, do_normalize=False, do_rescale=False, num_labels=self.image_processor_tester.num_classes
)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test whether the method "pad_and_return_pixel_mask" and calling the feature extractor return the same tensors
encoded_images_with_method = feature_extractor_1.encode_inputs(image_inputs, return_tensors="pt")
encoded_images = feature_extractor_2(image_inputs, return_tensors="pt")
# Test whether the method "pad_and_return_pixel_mask" and calling the image processor return the same tensors
encoded_images_with_method = image_processing_1.encode_inputs(image_inputs, return_tensors="pt")
encoded_images = image_processing_2(image_inputs, return_tensors="pt")
self.assertTrue(
torch.allclose(encoded_images_with_method["pixel_values"], encoded_images["pixel_values"], atol=1e-4)
@@ -287,15 +286,15 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
torch.allclose(encoded_images_with_method["pixel_mask"], encoded_images["pixel_mask"], atol=1e-4)
)
def comm_get_feature_extractor_inputs(
def comm_get_image_processing_inputs(
self, with_segmentation_maps=False, is_instance_map=False, segmentation_type="np"
):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
image_processing = self.image_processing_class(**self.image_processor_dict)
# prepare image and target
num_labels = self.feature_extract_tester.num_labels
num_labels = self.image_processor_tester.num_labels
annotations = None
instance_id_to_semantic_id = None
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
if with_segmentation_maps:
high = num_labels
if is_instance_map:
@@ -309,7 +308,7 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
if segmentation_type == "pil":
annotations = [Image.fromarray(annotation) for annotation in annotations]
inputs = feature_extractor(
inputs = image_processing(
image_inputs,
annotations,
return_tensors="pt",
@@ -326,10 +325,10 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
size_divisors = [8, 16, 32]
weird_input_sizes = [(407, 802), (582, 1094)]
for size_divisor in size_divisors:
feat_extract_dict = {**self.feat_extract_dict, **{"size_divisor": size_divisor}}
feature_extractor = self.feature_extraction_class(**feat_extract_dict)
image_processor_dict = {**self.image_processor_dict, **{"size_divisor": size_divisor}}
image_processing = self.image_processing_class(**image_processor_dict)
for weird_input_size in weird_input_sizes:
inputs = feature_extractor([np.ones((3, *weird_input_size))], return_tensors="pt")
inputs = image_processing([np.ones((3, *weird_input_size))], return_tensors="pt")
pixel_values = inputs["pixel_values"]
# check if divisible
self.assertTrue((pixel_values.shape[-1] % size_divisor) == 0)
@@ -337,7 +336,7 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
def test_call_with_segmentation_maps(self):
def common(is_instance_map=False, segmentation_type=None):
inputs = self.comm_get_feature_extractor_inputs(
inputs = self.comm_get_image_processing_inputs(
with_segmentation_maps=True, is_instance_map=is_instance_map, segmentation_type=segmentation_type
)
@@ -389,11 +388,11 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
instance_seg1, inst2class1 = get_instance_segmentation_and_mapping(annotation1)
instance_seg2, inst2class2 = get_instance_segmentation_and_mapping(annotation2)
# create a feature extractor
feature_extractor = MaskFormerFeatureExtractor(reduce_labels=True, ignore_index=255, size=(512, 512))
# create a image processor
image_processing = MaskFormerImageProcessor(reduce_labels=True, ignore_index=255, size=(512, 512))
# prepare the images and annotations
inputs = feature_extractor(
inputs = image_processing(
[image1, image2],
[instance_seg1, instance_seg2],
instance_id_to_semantic_id=[inst2class1, inst2class2],
@@ -432,11 +431,11 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
hf_hub_download(repo_id=repo_id, filename="semantic_segmentation_annotation_2.png", repo_type="dataset")
)
# create a feature extractor
feature_extractor = MaskFormerFeatureExtractor(reduce_labels=True, ignore_index=255, size=(512, 512))
# create a image processor
image_processing = MaskFormerImageProcessor(reduce_labels=True, ignore_index=255, size=(512, 512))
# prepare the images and annotations
inputs = feature_extractor(
inputs = image_processing(
[image1, image2],
[annotation1, annotation2],
return_tensors="pt",
@@ -489,12 +488,12 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
panoptic_map1, inst2class1 = create_panoptic_map(annotation1, segments_info1)
panoptic_map2, inst2class2 = create_panoptic_map(annotation2, segments_info2)
# create a feature extractor
feature_extractor = MaskFormerFeatureExtractor(ignore_index=0, do_resize=False)
# create a image processor
image_processing = MaskFormerImageProcessor(ignore_index=0, do_resize=False)
# prepare the images and annotations
pixel_values_list = [np.moveaxis(np.array(image1), -1, 0), np.moveaxis(np.array(image2), -1, 0)]
inputs = feature_extractor.encode_inputs(
inputs = image_processing.encode_inputs(
pixel_values_list,
[panoptic_map1, panoptic_map2],
instance_id_to_semantic_id=[inst2class1, inst2class2],
@@ -535,17 +534,17 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
self.assertEqual(rle[1], 45)
def test_post_process_segmentation(self):
fature_extractor = self.feature_extraction_class(num_labels=self.feature_extract_tester.num_classes)
outputs = self.feature_extract_tester.get_fake_maskformer_outputs()
fature_extractor = self.image_processing_class(num_labels=self.image_processor_tester.num_classes)
outputs = self.image_processor_tester.get_fake_maskformer_outputs()
segmentation = fature_extractor.post_process_segmentation(outputs)
self.assertEqual(
segmentation.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_classes,
self.feature_extract_tester.height,
self.feature_extract_tester.width,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_classes,
self.image_processor_tester.height,
self.image_processor_tester.width,
),
)
@@ -554,53 +553,53 @@ class MaskFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
self.assertEqual(
segmentation.shape,
(self.feature_extract_tester.batch_size, self.feature_extract_tester.num_classes, *target_size),
(self.image_processor_tester.batch_size, self.image_processor_tester.num_classes, *target_size),
)
def test_post_process_semantic_segmentation(self):
fature_extractor = self.feature_extraction_class(num_labels=self.feature_extract_tester.num_classes)
outputs = self.feature_extract_tester.get_fake_maskformer_outputs()
fature_extractor = self.image_processing_class(num_labels=self.image_processor_tester.num_classes)
outputs = self.image_processor_tester.get_fake_maskformer_outputs()
segmentation = fature_extractor.post_process_semantic_segmentation(outputs)
self.assertEqual(len(segmentation), self.feature_extract_tester.batch_size)
self.assertEqual(len(segmentation), self.image_processor_tester.batch_size)
self.assertEqual(
segmentation[0].shape,
(
self.feature_extract_tester.height,
self.feature_extract_tester.width,
self.image_processor_tester.height,
self.image_processor_tester.width,
),
)
target_sizes = [(1, 4) for i in range(self.feature_extract_tester.batch_size)]
target_sizes = [(1, 4) for i in range(self.image_processor_tester.batch_size)]
segmentation = fature_extractor.post_process_semantic_segmentation(outputs, target_sizes=target_sizes)
self.assertEqual(segmentation[0].shape, target_sizes[0])
def test_post_process_panoptic_segmentation(self):
feature_extractor = self.feature_extraction_class(num_labels=self.feature_extract_tester.num_classes)
outputs = self.feature_extract_tester.get_fake_maskformer_outputs()
segmentation = feature_extractor.post_process_panoptic_segmentation(outputs, threshold=0)
image_processing = self.image_processing_class(num_labels=self.image_processor_tester.num_classes)
outputs = self.image_processor_tester.get_fake_maskformer_outputs()
segmentation = image_processing.post_process_panoptic_segmentation(outputs, threshold=0)
self.assertTrue(len(segmentation) == self.feature_extract_tester.batch_size)
self.assertTrue(len(segmentation) == self.image_processor_tester.batch_size)
for el in segmentation:
self.assertTrue("segmentation" in el)
self.assertTrue("segments_info" in el)
self.assertEqual(type(el["segments_info"]), list)
self.assertEqual(
el["segmentation"].shape, (self.feature_extract_tester.height, self.feature_extract_tester.width)
el["segmentation"].shape, (self.image_processor_tester.height, self.image_processor_tester.width)
)
def test_post_process_label_fusing(self):
feature_extractor = self.feature_extraction_class(num_labels=self.feature_extract_tester.num_classes)
outputs = self.feature_extract_tester.get_fake_maskformer_outputs()
image_processor = self.image_processing_class(num_labels=self.image_processor_tester.num_classes)
outputs = self.image_processor_tester.get_fake_maskformer_outputs()
segmentation = feature_extractor.post_process_panoptic_segmentation(
segmentation = image_processor.post_process_panoptic_segmentation(
outputs, threshold=0, mask_threshold=0, overlap_mask_area_threshold=0
)
unfused_segments = [el["segments_info"] for el in segmentation]
fused_segmentation = feature_extractor.post_process_panoptic_segmentation(
fused_segmentation = image_processor.post_process_panoptic_segmentation(
outputs, threshold=0, mask_threshold=0, overlap_mask_area_threshold=0, label_ids_to_fuse={1}
)
fused_segments = [el["segments_info"] for el in fused_segmentation]

117
tests/models/mobilenet_v1/test_image_processing_mobilenet_v1.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import MobileNetV1FeatureExtractor
from transformers import MobileNetV1ImageProcessor
class MobileNetV1FeatureExtractionTester(unittest.TestCase):
class MobileNetV1ImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -61,7 +60,7 @@ class MobileNetV1FeatureExtractionTester(unittest.TestCase):
self.do_center_crop = do_center_crop
self.crop_size = crop_size
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -72,128 +71,128 @@ class MobileNetV1FeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class MobileNetV1FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class MobileNetV1ImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = MobileNetV1FeatureExtractor if is_vision_available() else None
image_processing_class = MobileNetV1ImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = MobileNetV1FeatureExtractionTester(self)
self.image_processor_tester = MobileNetV1ImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 20})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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": 20})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

117
tests/models/mobilenet_v2/test_image_processing_mobilenet_v2.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import MobileNetV2FeatureExtractor
from transformers import MobileNetV2ImageProcessor
class MobileNetV2FeatureExtractionTester(unittest.TestCase):
class MobileNetV2ImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -61,7 +60,7 @@ class MobileNetV2FeatureExtractionTester(unittest.TestCase):
self.do_center_crop = do_center_crop
self.crop_size = crop_size
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -72,128 +71,128 @@ class MobileNetV2FeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class MobileNetV2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class MobileNetV2ImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = MobileNetV2FeatureExtractor if is_vision_available() else None
image_processing_class = MobileNetV2ImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = MobileNetV2FeatureExtractionTester(self)
self.image_processor_tester = MobileNetV2ImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "crop_size"))
def test_image_processor_properties(self):
image_processor = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processor, "do_resize"))
self.assertTrue(hasattr(image_processor, "size"))
self.assertTrue(hasattr(image_processor, "do_center_crop"))
self.assertTrue(hasattr(image_processor, "crop_size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 20})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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": 20})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

119
tests/models/mobilevit/test_image_processing_mobilevit.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import MobileViTFeatureExtractor
from transformers import MobileViTImageProcessor
class MobileViTFeatureExtractionTester(unittest.TestCase):
class MobileViTImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -63,7 +62,7 @@ class MobileViTFeatureExtractionTester(unittest.TestCase):
self.crop_size = crop_size
self.do_flip_channel_order = do_flip_channel_order
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -75,129 +74,129 @@ class MobileViTFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class MobileViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class MobileViTImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = MobileViTFeatureExtractor if is_vision_available() else None
image_processing_class = MobileViTImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = MobileViTFeatureExtractionTester(self)
self.image_processor_tester = MobileViTImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_flip_channel_order"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_flip_channel_order"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 20})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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": 20})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

29
tests/models/oneformer/test_image_processing_oneformer.py
View File

@@ -23,8 +23,7 @@ from huggingface_hub import hf_hub_download
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -100,7 +99,7 @@ class OneFormerImageProcessorTester(unittest.TestCase):
self.do_reduce_labels = do_reduce_labels
self.ignore_index = ignore_index
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -156,20 +155,20 @@ class OneFormerImageProcessorTester(unittest.TestCase):
@require_torch
@require_vision
class OneFormerImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class OneFormerImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
image_processing_class = OneFormerImageProcessor if (is_vision_available() and is_torch_available()) else None
# only for test_feat_extracttion_common.test_feat_extract_to_json_string
feature_extraction_class = image_processing_class
# only for test_image_processing_common.test_image_proc_to_json_string
image_processing_class = image_processing_class
def setUp(self):
self.image_processing_tester = OneFormerImageProcessorTester(self)
@property
def feat_extract_dict(self):
return self.image_processing_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processing_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
image_processor = self.image_processing_class(**self.feat_extract_dict)
def test_image_proc_properties(self):
image_processor = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processor, "image_mean"))
self.assertTrue(hasattr(image_processor, "image_std"))
self.assertTrue(hasattr(image_processor, "do_normalize"))
@@ -187,7 +186,7 @@ class OneFormerImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.Te
def test_call_pil(self):
# Initialize image_processor
image_processor = self.image_processing_class(**self.feat_extract_dict)
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.image_processing_tester, equal_resolution=False)
for image in image_inputs:
@@ -221,7 +220,7 @@ class OneFormerImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.Te
def test_call_numpy(self):
# Initialize image_processor
image_processor = self.image_processing_class(**self.feat_extract_dict)
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.image_processing_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
@@ -255,7 +254,7 @@ class OneFormerImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.Te
def test_call_pytorch(self):
# Initialize image_processor
image_processor = self.image_processing_class(**self.feat_extract_dict)
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.image_processing_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
@@ -289,7 +288,7 @@ class OneFormerImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.Te
def test_equivalence_pad_and_create_pixel_mask(self):
# Initialize image_processors
image_processor_1 = self.image_processing_class(**self.feat_extract_dict)
image_processor_1 = self.image_processing_class(**self.image_processor_dict)
image_processor_2 = self.image_processing_class(
do_resize=False,
do_normalize=False,
@@ -320,7 +319,7 @@ class OneFormerImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.Te
def comm_get_image_processor_inputs(
self, with_segmentation_maps=False, is_instance_map=False, segmentation_type="np"
):
image_processor = self.image_processing_class(**self.feat_extract_dict)
image_processor = self.image_processing_class(**self.image_processor_dict)
# prepare image and target
num_labels = self.image_processing_tester.num_labels
annotations = None

125
tests/models/owlvit/test_image_processing_owlvit.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import OwlViTFeatureExtractor
from transformers import OwlViTImageProcessor
class OwlViTFeatureExtractionTester(unittest.TestCase):
class OwlViTImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -67,7 +66,7 @@ class OwlViTFeatureExtractionTester(unittest.TestCase):
self.image_std = image_std
self.do_convert_rgb = do_convert_rgb
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -82,130 +81,130 @@ class OwlViTFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class OwlViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class OwlViTImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = OwlViTFeatureExtractor if is_vision_available() else None
image_processing_class = OwlViTImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = OwlViTFeatureExtractionTester(self)
self.image_processor_tester = OwlViTImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "center_crop"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_convert_rgb"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "center_crop"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_convert_rgb"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 18, "width": 18})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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, {"height": 18, "width": 18})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

121
tests/models/poolformer/test_image_processing_poolformer.py
View File

@@ -20,8 +20,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -30,10 +29,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import PoolFormerFeatureExtractor
from transformers import PoolFormerImageProcessor
class PoolFormerFeatureExtractionTester(unittest.TestCase):
class PoolFormerImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -64,7 +63,7 @@ class PoolFormerFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"size": self.size,
"do_resize_and_center_crop": self.do_resize_and_center_crop,
@@ -78,131 +77,131 @@ class PoolFormerFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class PoolFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class PoolFormerImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = PoolFormerFeatureExtractor if is_vision_available() else None
image_processing_class = PoolFormerImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = PoolFormerFeatureExtractionTester(self)
self.image_processor_tester = PoolFormerImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize_and_center_crop"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "crop_pct"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize_and_center_crop"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "crop_pct"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 30})
self.assertEqual(feature_extractor.crop_size, {"height": 30, "width": 30})
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": 30})
self.assertEqual(image_processor.crop_size, {"height": 30, "width": 30})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

191
tests/models/segformer/test_image_processing_segformer.py
View File

@@ -22,8 +22,7 @@ from datasets import load_dataset
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -32,10 +31,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import SegformerFeatureExtractor
from transformers import SegformerImageProcessor
class SegformerFeatureExtractionTester(unittest.TestCase):
class SegformerImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -63,7 +62,7 @@ class SegformerFeatureExtractionTester(unittest.TestCase):
self.image_std = image_std
self.do_reduce_labels = do_reduce_labels
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -96,163 +95,161 @@ def prepare_semantic_batch_inputs():
@require_torch
@require_vision
class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class SegformerImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = SegformerFeatureExtractor if is_vision_available() else None
image_processing_class = SegformerImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = SegformerFeatureExtractionTester(self)
self.image_processor_tester = SegformerImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_reduce_labels"))
def test_image_processor_properties(self):
image_processing = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processing, "do_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "do_normalize"))
self.assertTrue(hasattr(image_processing, "image_mean"))
self.assertTrue(hasattr(image_processing, "image_std"))
self.assertTrue(hasattr(image_processing, "do_reduce_labels"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 30, "width": 30})
self.assertEqual(feature_extractor.do_reduce_labels, False)
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, {"height": 30, "width": 30})
self.assertEqual(image_processor.do_reduce_labels, False)
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, reduce_labels=True
)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
self.assertEqual(feature_extractor.do_reduce_labels, True)
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, reduce_labels=True)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
self.assertEqual(image_processor.do_reduce_labels, True)
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_segmentation_maps(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
maps = []
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
maps.append(torch.zeros(image.shape[-2:]).long())
# Test not batched input
encoding = feature_extractor(image_inputs[0], maps[0], return_tensors="pt")
encoding = image_processing(image_inputs[0], maps[0], return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -260,22 +257,22 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
self.assertTrue(encoding["labels"].max().item() <= 255)
# Test batched
encoding = feature_extractor(image_inputs, maps, return_tensors="pt")
encoding = image_processing(image_inputs, maps, return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -285,22 +282,22 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
# Test not batched input (PIL images)
image, segmentation_map = prepare_semantic_single_inputs()
encoding = feature_extractor(image, segmentation_map, return_tensors="pt")
encoding = image_processing(image, segmentation_map, return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -310,22 +307,22 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
# Test batched input (PIL images)
images, segmentation_maps = prepare_semantic_batch_inputs()
encoding = feature_extractor(images, segmentation_maps, return_tensors="pt")
encoding = image_processing(images, segmentation_maps, return_tensors="pt")
self.assertEqual(
encoding["pixel_values"].shape,
(
2,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
2,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -333,16 +330,16 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
self.assertTrue(encoding["labels"].max().item() <= 255)
def test_reduce_labels(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# ADE20k has 150 classes, and the background is included, so labels should be between 0 and 150
image, map = prepare_semantic_single_inputs()
encoding = feature_extractor(image, map, return_tensors="pt")
encoding = image_processing(image, map, return_tensors="pt")
self.assertTrue(encoding["labels"].min().item() >= 0)
self.assertTrue(encoding["labels"].max().item() <= 150)
feature_extractor.reduce_labels = True
encoding = feature_extractor(image, map, return_tensors="pt")
image_processing.reduce_labels = True
encoding = image_processing(image, map, return_tensors="pt")
self.assertTrue(encoding["labels"].min().item() >= 0)
self.assertTrue(encoding["labels"].max().item() <= 255)

58
tests/models/swin2sr/test_image_processing_swin2sr.py
View File

@@ -21,7 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import ImageProcessingSavingTestMixin
if is_torch_available():
@@ -59,7 +59,7 @@ class Swin2SRImageProcessingTester(unittest.TestCase):
self.do_pad = do_pad
self.pad_size = pad_size
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_rescale": self.do_rescale,
"rescale_factor": self.rescale_factor,
@@ -100,93 +100,93 @@ class Swin2SRImageProcessingTester(unittest.TestCase):
@require_torch
@require_vision
class Swin2SRImageProcessingTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class Swin2SRImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = Swin2SRImageProcessor if is_vision_available() else None
image_processing_class = Swin2SRImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = Swin2SRImageProcessingTester(self)
self.image_processor_tester = Swin2SRImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "do_rescale"))
self.assertTrue(hasattr(feature_extractor, "rescale_factor"))
self.assertTrue(hasattr(feature_extractor, "do_pad"))
self.assertTrue(hasattr(feature_extractor, "pad_size"))
def test_image_processor_properties(self):
image_processor = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(image_processor, "do_rescale"))
self.assertTrue(hasattr(image_processor, "rescale_factor"))
self.assertTrue(hasattr(image_processor, "do_pad"))
self.assertTrue(hasattr(image_processor, "pad_size"))
def test_batch_feature(self):
pass
def calculate_expected_size(self, image):
old_height, old_width = get_image_size(image)
size = self.feature_extract_tester.pad_size
size = self.image_processor_tester.pad_size
pad_height = (old_height // size + 1) * size - old_height
pad_width = (old_width // size + 1) * size - old_width
return old_height + pad_height, old_width + pad_width
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.calculate_expected_size(np.array(image_inputs[0]))
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, numpify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.calculate_expected_size(image_inputs[0])
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processor
image_processor = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = self.feature_extract_tester.prepare_inputs(equal_resolution=False, torchify=True)
image_inputs = self.image_processor_tester.prepare_inputs(equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processor(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.calculate_expected_size(image_inputs[0])
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),

133
tests/models/videomae/test_image_processing_videomae.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_video_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_video_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import VideoMAEFeatureExtractor
from transformers import VideoMAEImageProcessor
class VideoMAEFeatureExtractionTester(unittest.TestCase):
class VideoMAEImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -68,7 +67,7 @@ class VideoMAEFeatureExtractionTester(unittest.TestCase):
self.image_std = image_std
self.crop_size = crop_size
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -81,139 +80,139 @@ class VideoMAEFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class VideoMAEImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = VideoMAEFeatureExtractor if is_vision_available() else None
image_processing_class = VideoMAEImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = VideoMAEFeatureExtractionTester(self)
self.image_processor_tester = VideoMAEImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "do_center_crop"))
self.assertTrue(hasattr(feature_extractor, "size"))
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_resize"))
self.assertTrue(hasattr(image_processing, "do_center_crop"))
self.assertTrue(hasattr(image_processing, "size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 18})
self.assertEqual(feature_extractor.crop_size, {"height": 18, "width": 18})
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})
self.assertEqual(image_processor.crop_size, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42, crop_size=84)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
self.assertEqual(feature_extractor.crop_size, {"height": 84, "width": 84})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42, crop_size=84)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
self.assertEqual(image_processor.crop_size, {"height": 84, "width": 84})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL videos
video_inputs = prepare_video_inputs(self.feature_extract_tester, equal_resolution=False)
video_inputs = prepare_video_inputs(self.image_processor_tester, equal_resolution=False)
for video in video_inputs:
self.assertIsInstance(video, list)
self.assertIsInstance(video[0], Image.Image)
# Test not batched input
encoded_videos = feature_extractor(video_inputs[0], return_tensors="pt").pixel_values
encoded_videos = image_processing(video_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_videos.shape,
(
1,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_frames,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_videos = feature_extractor(video_inputs, return_tensors="pt").pixel_values
encoded_videos = image_processing(video_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_videos.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_frames,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
video_inputs = prepare_video_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
video_inputs = prepare_video_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for video in video_inputs:
self.assertIsInstance(video, list)
self.assertIsInstance(video[0], np.ndarray)
# Test not batched input
encoded_videos = feature_extractor(video_inputs[0], return_tensors="pt").pixel_values
encoded_videos = image_processing(video_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_videos.shape,
(
1,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_frames,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_videos = feature_extractor(video_inputs, return_tensors="pt").pixel_values
encoded_videos = image_processing(video_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_videos.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_frames,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
video_inputs = prepare_video_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
video_inputs = prepare_video_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for video in video_inputs:
self.assertIsInstance(video, list)
self.assertIsInstance(video[0], torch.Tensor)
# Test not batched input
encoded_videos = feature_extractor(video_inputs[0], return_tensors="pt").pixel_values
encoded_videos = image_processing(video_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_videos.shape,
(
1,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.num_frames,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)
# Test batched
encoded_videos = feature_extractor(video_inputs, return_tensors="pt").pixel_values
encoded_videos = image_processing(video_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_videos.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_frames,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
),
)

121
tests/models/vilt/test_image_processing_vilt.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import ViltFeatureExtractor
from transformers import ViltImageProcessor
class ViltFeatureExtractionTester(unittest.TestCase):
class ViltImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -64,7 +63,7 @@ class ViltFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -76,7 +75,7 @@ class ViltFeatureExtractionTester(unittest.TestCase):
def get_expected_values(self, image_inputs, batched=False):
"""
This function computes the expected height and width when providing images to ViltFeatureExtractor,
This function computes the expected height and width when providing images to ViltImageProcessor,
assuming do_resize is set to True with a scalar size and size_divisor.
"""
if not batched:
@@ -117,141 +116,141 @@ class ViltFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class ViltFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class ViltImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ViltFeatureExtractor if is_vision_available() else None
image_processing_class = ViltImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = ViltFeatureExtractionTester(self)
self.image_processor_tester = ViltImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
self.assertTrue(hasattr(feature_extractor, "size_divisor"))
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_resize"))
self.assertTrue(hasattr(image_processing, "size"))
self.assertTrue(hasattr(image_processing, "size_divisor"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 30})
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": 30})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"shortest_edge": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"shortest_edge": 42})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_equivalence_pad_and_create_pixel_mask(self):
# Initialize feature_extractors
feature_extractor_1 = self.feature_extraction_class(**self.feat_extract_dict)
feature_extractor_2 = self.feature_extraction_class(do_resize=False, do_normalize=False, do_rescale=False)
# Initialize image_processings
image_processing_1 = self.image_processing_class(**self.image_processor_dict)
image_processing_2 = self.image_processing_class(do_resize=False, do_normalize=False, do_rescale=False)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test whether the method "pad_and_return_pixel_mask" and calling the feature extractor return the same tensors
encoded_images_with_method = feature_extractor_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = feature_extractor_2(image_inputs, return_tensors="pt")
# Test whether the method "pad_and_return_pixel_mask" and calling the image processor return the same tensors
encoded_images_with_method = image_processing_1.pad_and_create_pixel_mask(image_inputs, return_tensors="pt")
encoded_images = image_processing_2(image_inputs, return_tensors="pt")
self.assertTrue(
torch.allclose(encoded_images_with_method["pixel_values"], encoded_images["pixel_values"], atol=1e-4)

115
tests/models/vit/test_image_processing_vit.py
View File

@@ -21,8 +21,7 @@ import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -31,10 +30,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import ViTFeatureExtractor
from transformers import ViTImageProcessor
class ViTFeatureExtractionTester(unittest.TestCase):
class ViTImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -62,7 +61,7 @@ class ViTFeatureExtractionTester(unittest.TestCase):
self.image_mean = image_mean
self.image_std = image_std
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -74,127 +73,127 @@ class ViTFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class ViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class ViTImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = ViTFeatureExtractor if is_vision_available() else None
image_processing_class = ViTImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = ViTFeatureExtractionTester(self)
self.image_processor_tester = ViTImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
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_resize"))
self.assertTrue(hasattr(image_processing, "size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"height": 18, "width": 18})
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, {"height": 18, "width": 18})
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict, size=42)
self.assertEqual(feature_extractor.size, {"height": 42, "width": 42})
image_processor = self.image_processing_class.from_dict(self.image_processor_dict, size=42)
self.assertEqual(image_processor.size, {"height": 42, "width": 42})
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size["height"],
self.image_processor_tester.size["width"],
),
)

133
tests/models/yolos/test_image_processing_yolos.py
View File

@@ -23,8 +23,7 @@ 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 ...test_feature_extraction_common import FeatureExtractionSavingTestMixin
from ...test_image_processing_common import prepare_image_inputs
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
@@ -33,10 +32,10 @@ if is_torch_available():
if is_vision_available():
from PIL import Image
from transformers import YolosFeatureExtractor
from transformers import YolosImageProcessor
class YolosFeatureExtractionTester(unittest.TestCase):
class YolosImageProcessingTester(unittest.TestCase):
def __init__(
self,
parent,
@@ -69,7 +68,7 @@ class YolosFeatureExtractionTester(unittest.TestCase):
self.rescale_factor = rescale_factor
self.do_pad = do_pad
def prepare_feat_extract_dict(self):
def prepare_image_processor_dict(self):
return {
"do_resize": self.do_resize,
"size": self.size,
@@ -83,7 +82,7 @@ class YolosFeatureExtractionTester(unittest.TestCase):
def get_expected_values(self, image_inputs, batched=False):
"""
This function computes the expected height and width when providing images to YolosFeatureExtractor,
This function computes the expected height and width when providing images to YolosImageProcessor,
assuming do_resize is set to True with a scalar size.
"""
if not batched:
@@ -115,149 +114,149 @@ class YolosFeatureExtractionTester(unittest.TestCase):
@require_torch
@require_vision
class YolosFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCase):
class YolosImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
feature_extraction_class = YolosFeatureExtractor if is_vision_available() else None
image_processing_class = YolosImageProcessor if is_vision_available() else None
def setUp(self):
self.feature_extract_tester = YolosFeatureExtractionTester(self)
self.image_processor_tester = YolosImageProcessingTester(self)
@property
def feat_extract_dict(self):
return self.feature_extract_tester.prepare_feat_extract_dict()
def image_processor_dict(self):
return self.image_processor_tester.prepare_image_processor_dict()
def test_feat_extract_properties(self):
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
self.assertTrue(hasattr(feature_extractor, "image_mean"))
self.assertTrue(hasattr(feature_extractor, "image_std"))
self.assertTrue(hasattr(feature_extractor, "do_normalize"))
self.assertTrue(hasattr(feature_extractor, "do_resize"))
self.assertTrue(hasattr(feature_extractor, "size"))
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_resize"))
self.assertTrue(hasattr(image_processing, "size"))
def test_feat_extract_from_dict_with_kwargs(self):
feature_extractor = self.feature_extraction_class.from_dict(self.feat_extract_dict)
self.assertEqual(feature_extractor.size, {"shortest_edge": 18, "longest_edge": 1333})
self.assertEqual(feature_extractor.do_pad, True)
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)
feature_extractor = self.feature_extraction_class.from_dict(
self.feat_extract_dict, size=42, max_size=84, pad_and_return_pixel_mask=False
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(feature_extractor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(feature_extractor.do_pad, False)
self.assertEqual(image_processor.size, {"shortest_edge": 42, "longest_edge": 84})
self.assertEqual(image_processor.do_pad, False)
def test_batch_feature(self):
pass
def test_call_pil(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_numpy(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, numpify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, numpify=True)
for image in image_inputs:
self.assertIsInstance(image, np.ndarray)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_call_pytorch(self):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
# Initialize image_processing
image_processing = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs)
self.assertEqual(
encoded_images.shape,
(1, self.feature_extract_tester.num_channels, expected_height, expected_width),
(1, self.image_processor_tester.num_channels, expected_height, expected_width),
)
# Test batched
encoded_images = feature_extractor(image_inputs, return_tensors="pt").pixel_values
encoded_images = image_processing(image_inputs, return_tensors="pt").pixel_values
expected_height, expected_width = self.feature_extract_tester.get_expected_values(image_inputs, batched=True)
expected_height, expected_width = self.image_processor_tester.get_expected_values(image_inputs, batched=True)
self.assertEqual(
encoded_images.shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
),
)
def test_equivalence_padding(self):
# Initialize feature_extractors
feature_extractor_1 = self.feature_extraction_class(**self.feat_extract_dict)
feature_extractor_2 = self.feature_extraction_class(do_resize=False, do_normalize=False, do_rescale=False)
# Initialize image_processings
image_processing_1 = self.image_processing_class(**self.image_processor_dict)
image_processing_2 = self.image_processing_class(do_resize=False, do_normalize=False, do_rescale=False)
# create random PyTorch tensors
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False, torchify=True)
image_inputs = prepare_image_inputs(self.image_processor_tester, equal_resolution=False, torchify=True)
for image in image_inputs:
self.assertIsInstance(image, torch.Tensor)
# Test whether the method "pad" and calling the feature extractor return the same tensors
encoded_images_with_method = feature_extractor_1.pad(image_inputs, return_tensors="pt")
encoded_images = feature_extractor_2(image_inputs, return_tensors="pt")
# Test whether the method "pad" and calling the image processor return the same tensors
encoded_images_with_method = image_processing_1.pad(image_inputs, return_tensors="pt")
encoded_images = image_processing_2(image_inputs, return_tensors="pt")
self.assertTrue(
torch.allclose(encoded_images_with_method["pixel_values"], encoded_images["pixel_values"], atol=1e-4)
@@ -273,8 +272,8 @@ class YolosFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
target = {"image_id": 39769, "annotations": target}
# encode them
feature_extractor = YolosFeatureExtractor.from_pretrained("hustvl/yolos-small")
encoding = feature_extractor(images=image, annotations=target, return_tensors="pt")
image_processing = YolosImageProcessor.from_pretrained("hustvl/yolos-small")
encoding = image_processing(images=image, annotations=target, return_tensors="pt")
# verify pixel values
expected_shape = torch.Size([1, 3, 800, 1066])
@@ -319,8 +318,8 @@ class YolosFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
# encode them
feature_extractor = YolosFeatureExtractor(format="coco_panoptic")
encoding = feature_extractor(images=image, annotations=target, masks_path=masks_path, return_tensors="pt")
image_processing = YolosImageProcessor(format="coco_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])