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

Add Image Processors (#19796)

Browse Source 
* Add CLIP image processor

* Crop size as dict too

* Update warning

* Actually use logger this time

* Normalize doesn't change dtype of input

* Add perceiver image processor

* Tidy up

* Add DPT image processor

* Add Vilt image processor

* Tidy up

* Add poolformer image processor

* Tidy up

* Add LayoutLM v2 and v3 imsge processors

* Tidy up

* Add Flava image processor

* Tidy up

* Add deit image processor

* Tidy up

* Add ConvNext image processor

* Tidy up

* Add levit image processor

* Add segformer image processor

* Add in post processing

* Fix up

* Add ImageGPT image processor

* Fixup

* Add mobilevit image processor

* Tidy up

* Add postprocessing

* Fixup

* Add VideoMAE image processor

* Tidy up

* Add ImageGPT image processor

* Fixup

* Add ViT image processor

* Tidy up

* Add beit image processor

* Add mobilevit image processor

* Tidy up

* Add postprocessing

* Fixup

* Fix up

* Fix flava and remove tree module

* Fix image classification pipeline failing tests

* Update feature extractor in trainer scripts

* Update pad_if_smaller to accept tuple and int size

* Update for image segmentation pipeline

* Update src/transformers/models/perceiver/image_processing_perceiver.py

Co-authored-by: Alara Dirik <8944735+alaradirik@users.noreply.github.com>

* Update src/transformers/image_processing_utils.py

Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* Update src/transformers/models/beit/image_processing_beit.py

Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* PR comments - docstrings; remove accidentally added resize; var names

* Update docstrings

* Add exception if size is not in the right format

* Fix exception check

* Fix up

* Use shortest_edge in tuple in script

Co-authored-by: Alara Dirik <8944735+alaradirik@users.noreply.github.com>
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>
This commit is contained in:
amyeroberts
2022-11-02 11:57:36 +00:00
committed by GitHub
parent 2e3452af0f
commit a6b7759880
65 changed files with 7060 additions and 3590 deletions
Download Patch File Download Diff File Expand all files Collapse all files

68
tests/models/beit/test_feature_extraction_beit.py
View File

@@ -44,14 +44,16 @@ class BeitFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=20,
size=None,
do_center_crop=True,
crop_size=18,
crop_size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
reduce_labels=False,
do_reduce_labels=False,
):
size = size if size is not None else {"height": 20, "width": 20}
crop_size = crop_size if crop_size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -65,7 +67,7 @@ class BeitFeatureExtractionTester(unittest.TestCase):
self.do_normalize = do_normalize
self.image_mean = image_mean
self.image_std = image_std
self.reduce_labels = reduce_labels
self.do_reduce_labels = do_reduce_labels
def prepare_feat_extract_dict(self):
return {
@@ -76,7 +78,7 @@ class BeitFeatureExtractionTester(unittest.TestCase):
"do_normalize": self.do_normalize,
"image_mean": self.image_mean,
"image_std": self.image_std,
"reduce_labels": self.reduce_labels,
"do_reduce_labels": self.do_reduce_labels,
}
@@ -141,8 +143,8 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -153,8 +155,8 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -173,8 +175,8 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -185,8 +187,8 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -205,8 +207,8 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -217,8 +219,8 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -239,16 +241,16 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -262,16 +264,16 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -287,16 +289,16 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -312,16 +314,16 @@ class BeitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
2,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
2,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)

38
tests/models/clip/test_feature_extraction_clip.py
View File

@@ -43,14 +43,16 @@ class CLIPFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=20,
size=None,
do_center_crop=True,
crop_size=18,
crop_size=None,
do_normalize=True,
image_mean=[0.48145466, 0.4578275, 0.40821073],
image_std=[0.26862954, 0.26130258, 0.27577711],
do_convert_rgb=True,
):
size = size if size is not None else {"shortest_edge": 20}
crop_size = crop_size if crop_size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -151,8 +153,8 @@ class CLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -163,8 +165,8 @@ class CLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -183,8 +185,8 @@ class CLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -195,8 +197,8 @@ class CLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -215,8 +217,8 @@ class CLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -227,8 +229,8 @@ class CLIPFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -276,8 +278,8 @@ class CLIPFeatureExtractionTestFourChannels(FeatureExtractionSavingTestMixin, un
(
1,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -288,7 +290,7 @@ class CLIPFeatureExtractionTestFourChannels(FeatureExtractionSavingTestMixin, un
(
self.feature_extract_tester.batch_size,
self.expected_encoded_image_num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)

27
tests/models/convnext/test_feature_extraction_convnext.py
View File

@@ -43,12 +43,13 @@ class ConvNextFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=20,
size=None,
crop_pct=0.875,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
):
size = size if size is not None else {"shortest_edge": 20}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -113,8 +114,8 @@ class ConvNextFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
),
)
@@ -125,8 +126,8 @@ class ConvNextFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
),
)
@@ -145,8 +146,8 @@ class ConvNextFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
),
)
@@ -157,8 +158,8 @@ class ConvNextFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
),
)
@@ -177,8 +178,8 @@ class ConvNextFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
),
)
@@ -189,7 +190,7 @@ class ConvNextFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["shortest_edge"],
self.feature_extract_tester.size["shortest_edge"],
),
)

31
tests/models/deit/test_feature_extraction_deit.py
View File

@@ -43,13 +43,16 @@ class DeiTFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=20,
size=None,
do_center_crop=True,
crop_size=18,
crop_size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
):
size = size if size is not None else {"height": 20, "width": 20}
crop_size = crop_size if crop_size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -117,8 +120,8 @@ class DeiTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -129,8 +132,8 @@ class DeiTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -149,8 +152,8 @@ class DeiTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -161,8 +164,8 @@ class DeiTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -181,8 +184,8 @@ class DeiTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -193,7 +196,7 @@ class DeiTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)

27
tests/models/dpt/test_feature_extraction_dpt.py
View File

@@ -43,11 +43,12 @@ class DPTFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=18,
size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
):
size = size if size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -106,8 +107,8 @@ class DPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -118,8 +119,8 @@ class DPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -138,8 +139,8 @@ class DPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -150,8 +151,8 @@ class DPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -170,8 +171,8 @@ class DPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -182,7 +183,7 @@ class DPTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)

41
tests/models/flava/test_feature_extraction_flava.py
View File

@@ -28,11 +28,10 @@ if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
import PIL
from transformers import FlavaFeatureExtractor
from transformers.image_utils import PILImageResampling
from transformers.models.flava.feature_extraction_flava import (
from transformers.models.flava.image_processing_flava import (
FLAVA_CODEBOOK_MEAN,
FLAVA_CODEBOOK_STD,
FLAVA_IMAGE_MEAN,
@@ -51,10 +50,12 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=224,
size=None,
do_center_crop=True,
crop_size=224,
crop_size=None,
resample=None,
do_rescale=True,
rescale_factor=1 / 255,
do_normalize=True,
image_mean=FLAVA_IMAGE_MEAN,
image_std=FLAVA_IMAGE_STD,
@@ -65,23 +66,30 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
mask_group_min_aspect_ratio=0.3,
mask_group_max_aspect_ratio=None,
codebook_do_resize=True,
codebook_size=112,
codebook_size=None,
codebook_resample=None,
codebook_do_center_crop=True,
codebook_crop_size=112,
codebook_crop_size=None,
codebook_do_map_pixels=True,
codebook_do_normalize=True,
codebook_image_mean=FLAVA_CODEBOOK_MEAN,
codebook_image_std=FLAVA_CODEBOOK_STD,
):
size = size if size is not None else {"height": 224, "width": 224}
crop_size = crop_size if crop_size is not None else {"height": 224, "width": 224}
codebook_size = codebook_size if codebook_size is not None else {"height": 112, "width": 112}
codebook_crop_size = codebook_crop_size if codebook_crop_size is not None else {"height": 112, "width": 112}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
self.do_resize = do_resize
self.do_rescale = do_rescale
self.rescale_factor = rescale_factor
self.min_resolution = min_resolution
self.max_resolution = max_resolution
self.size = size
self.resample = resample if resample is not None else PILImageResampling.BICUBIC
self.resample = resample if resample is not None else PIL.Image.Resampling.BICUBIC
self.do_normalize = do_normalize
self.image_mean = image_mean
self.image_std = image_std
@@ -97,7 +105,7 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
self.codebook_do_resize = codebook_do_resize
self.codebook_size = codebook_size
self.codebook_resample = codebook_resample if codebook_resample is not None else PILImageResampling.LANCZOS
self.codebook_resample = codebook_resample if codebook_resample is not None else PIL.Image.Resampling.LANCZOS
self.codebook_do_center_crop = codebook_do_center_crop
self.codebook_crop_size = codebook_crop_size
self.codebook_do_map_pixels = codebook_do_map_pixels
@@ -113,6 +121,8 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
"do_resize": self.do_resize,
"size": self.size,
"resample": self.resample,
"do_rescale": self.do_rescale,
"rescale_factor": self.rescale_factor,
"do_center_crop": self.do_center_crop,
"crop_size": self.crop_size,
"input_size_patches": self.input_size_patches,
@@ -133,7 +143,7 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
}
def get_expected_image_size(self):
return (self.size, self.size) if not isinstance(self.size, tuple) else self.size
return (self.size["height"], self.size["width"])
def get_expected_mask_size(self):
return (
@@ -143,10 +153,7 @@ class FlavaFeatureExtractionTester(unittest.TestCase):
)
def get_expected_codebook_image_size(self):
if not isinstance(self.codebook_size, tuple):
return (self.codebook_size, self.codebook_size)
else:
return self.codebook_size
return (self.codebook_size["height"], self.codebook_size["width"])
@require_torch
@@ -172,6 +179,8 @@ class FlavaFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
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"))
@@ -192,7 +201,7 @@ class FlavaFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
self.assertIsInstance(image, PIL.Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_tensors="pt")
@@ -324,7 +333,7 @@ class FlavaFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)
for image in image_inputs:
self.assertIsInstance(image, Image.Image)
self.assertIsInstance(image, PIL.Image.Image)
# Test not batched input
encoded_images = feature_extractor(image_inputs[0], return_codebook_pixels=True, return_tensors="pt")

3
tests/models/flava/test_processor_flava.py
View File

@@ -32,7 +32,7 @@ if is_vision_available():
from PIL import Image
from transformers import FlavaFeatureExtractor, FlavaProcessor
from transformers.models.flava.feature_extraction_flava import (
from transformers.models.flava.image_processing_flava import (
FLAVA_CODEBOOK_MEAN,
FLAVA_CODEBOOK_STD,
FLAVA_IMAGE_MEAN,
@@ -69,7 +69,6 @@ class FlavaProcessorTest(unittest.TestCase):
"mask_group_max_aspect_ratio": None,
"codebook_do_resize": True,
"codebook_size": 112,
"codebook_resample": None,
"codebook_do_center_crop": True,
"codebook_crop_size": 112,
"codebook_do_map_pixels": True,

3
tests/models/imagegpt/test_feature_extraction_imagegpt.py
View File

@@ -47,9 +47,10 @@ class ImageGPTFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=18,
size=None,
do_normalize=True,
):
size = size if size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels

37
tests/models/layoutlmv2/test_feature_extraction_layoutlmv2.py
View File

@@ -43,9 +43,10 @@ class LayoutLMv2FeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=18,
size=None,
apply_ocr=True,
):
size = size if size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -97,8 +98,8 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -112,8 +113,8 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -132,8 +133,8 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -144,8 +145,8 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -164,8 +165,8 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -176,8 +177,8 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -210,12 +211,4 @@ class LayoutLMv2FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
encoding = feature_extractor(image, return_tensors="pt")
self.assertEqual(
encoding.pixel_values.shape,
(
1,
3,
224,
224,
),
)
self.assertEqual(encoding.pixel_values.shape, (1, 3, 224, 224))

27
tests/models/layoutlmv3/test_feature_extraction_layoutlmv3.py
View File

@@ -43,9 +43,10 @@ class LayoutLMv3FeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=18,
size=None,
apply_ocr=True,
):
size = size if size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -97,8 +98,8 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -112,8 +113,8 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -132,8 +133,8 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -144,8 +145,8 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -164,8 +165,8 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -176,8 +177,8 @@ class LayoutLMv3FeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)

31
tests/models/levit/test_feature_extraction_levit.py
View File

@@ -43,12 +43,15 @@ class LevitFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=18,
size=None,
do_center_crop=True,
crop_size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
):
size = size if size is not None else {"shortest_edge": 18}
crop_size = crop_size if crop_size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -58,6 +61,7 @@ class LevitFeatureExtractionTester(unittest.TestCase):
self.do_resize = do_resize
self.size = size
self.do_center_crop = do_center_crop
self.crop_size = crop_size
self.do_normalize = do_normalize
self.image_mean = image_mean
self.image_std = image_std
@@ -70,6 +74,7 @@ class LevitFeatureExtractionTester(unittest.TestCase):
"do_resize": self.do_resize,
"do_center_crop": self.do_center_crop,
"size": self.size,
"crop_size": self.crop_size,
}
@@ -113,8 +118,8 @@ class LevitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -125,8 +130,8 @@ class LevitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -145,8 +150,8 @@ class LevitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -157,8 +162,8 @@ class LevitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -177,8 +182,8 @@ class LevitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -189,7 +194,7 @@ class LevitFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.Test
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)

30
tests/models/mobilevit/test_feature_extraction_mobilevit.py
View File

@@ -43,11 +43,13 @@ class MobileViTFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=20,
size=None,
do_center_crop=True,
crop_size=18,
crop_size=None,
do_flip_channel_order=True,
):
size = size if size is not None else {"shortest_edge": 20}
crop_size = crop_size if crop_size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -109,8 +111,8 @@ class MobileViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -121,8 +123,8 @@ class MobileViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -141,8 +143,8 @@ class MobileViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -153,8 +155,8 @@ class MobileViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -173,8 +175,8 @@ class MobileViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -185,7 +187,7 @@ class MobileViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)

31
tests/models/poolformer/test_feature_extraction_poolformer.py
View File

@@ -41,12 +41,15 @@ class PoolFormerFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize_and_center_crop=True,
size=30,
size=None,
crop_pct=0.9,
crop_size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
):
size = size if size is not None else {"shortest_edge": 30}
crop_size = crop_size if crop_size is not None else {"height": 30, "width": 30}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -55,6 +58,7 @@ class PoolFormerFeatureExtractionTester(unittest.TestCase):
self.do_resize_and_center_crop = do_resize_and_center_crop
self.size = size
self.crop_pct = crop_pct
self.crop_size = crop_size
self.do_normalize = do_normalize
self.image_mean = image_mean
self.image_std = image_std
@@ -64,6 +68,7 @@ class PoolFormerFeatureExtractionTester(unittest.TestCase):
"size": self.size,
"do_resize_and_center_crop": self.do_resize_and_center_crop,
"crop_pct": self.crop_pct,
"crop_size": self.crop_size,
"do_normalize": self.do_normalize,
"image_mean": self.image_mean,
"image_std": self.image_std,
@@ -111,8 +116,8 @@ class PoolFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -123,8 +128,8 @@ class PoolFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -143,8 +148,8 @@ class PoolFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -155,8 +160,8 @@ class PoolFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -175,8 +180,8 @@ class PoolFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -187,7 +192,7 @@ class PoolFormerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)

67
tests/models/segformer/test_feature_extraction_segformer.py
View File

@@ -43,12 +43,13 @@ class SegformerFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=30,
size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
reduce_labels=False,
do_reduce_labels=False,
):
size = size if size is not None else {"height": 30, "width": 30}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -59,7 +60,7 @@ class SegformerFeatureExtractionTester(unittest.TestCase):
self.do_normalize = do_normalize
self.image_mean = image_mean
self.image_std = image_std
self.reduce_labels = reduce_labels
self.do_reduce_labels = do_reduce_labels
def prepare_feat_extract_dict(self):
return {
@@ -68,7 +69,7 @@ class SegformerFeatureExtractionTester(unittest.TestCase):
"do_normalize": self.do_normalize,
"image_mean": self.image_mean,
"image_std": self.image_std,
"reduce_labels": self.reduce_labels,
"do_reduce_labels": self.do_reduce_labels,
}
@@ -112,7 +113,7 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
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, "reduce_labels"))
self.assertTrue(hasattr(feature_extractor, "do_reduce_labels"))
def test_batch_feature(self):
pass
@@ -132,8 +133,8 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -144,8 +145,8 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -164,8 +165,8 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -176,8 +177,8 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -196,8 +197,8 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -208,8 +209,8 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -230,16 +231,16 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -253,16 +254,16 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -278,16 +279,16 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
1,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)
@@ -303,16 +304,16 @@ class SegformerFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.
(
2,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(
encoding["labels"].shape,
(
2,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
self.assertEqual(encoding["labels"].dtype, torch.long)

33
tests/models/videomae/test_feature_extraction_videomae.py
View File

@@ -44,11 +44,15 @@ class VideoMAEFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=18,
size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
crop_size=None,
):
size = size if size is not None else {"shortest_edge": 18}
crop_size = crop_size if crop_size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -61,6 +65,7 @@ class VideoMAEFeatureExtractionTester(unittest.TestCase):
self.do_normalize = do_normalize
self.image_mean = image_mean
self.image_std = image_std
self.crop_size = crop_size
def prepare_feat_extract_dict(self):
return {
@@ -69,6 +74,7 @@ class VideoMAEFeatureExtractionTester(unittest.TestCase):
"do_normalize": self.do_normalize,
"do_resize": self.do_resize,
"size": self.size,
"crop_size": self.crop_size,
}
@@ -91,6 +97,7 @@ class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
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_batch_feature(self):
@@ -113,8 +120,8 @@ class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
1,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -126,8 +133,8 @@ class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -148,8 +155,8 @@ class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
1,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -161,8 +168,8 @@ class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -183,8 +190,8 @@ class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
1,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)
@@ -196,7 +203,7 @@ class VideoMAEFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.T
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_frames,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.crop_size["height"],
self.feature_extract_tester.crop_size["width"],
),
)

14
tests/models/vilt/test_feature_extraction_vilt.py
View File

@@ -43,12 +43,13 @@ class ViltFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=30,
size=None,
size_divisor=2,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
):
size = size if size is not None else {"shortest_edge": 30}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -78,18 +79,19 @@ class ViltFeatureExtractionTester(unittest.TestCase):
assuming do_resize is set to True with a scalar size and size_divisor.
"""
if not batched:
size = self.size["shortest_edge"]
image = image_inputs[0]
if isinstance(image, Image.Image):
w, h = image.size
else:
h, w = image.shape[1], image.shape[2]
scale = self.size / min(w, h)
scale = size / min(w, h)
if h < w:
newh, neww = self.size, scale * w
newh, neww = size, scale * w
else:
newh, neww = scale * h, self.size
newh, neww = scale * h, size
max_size = int((1333 / 800) * self.size)
max_size = int((1333 / 800) * size)
if max(newh, neww) > max_size:
scale = max_size / max(newh, neww)
newh = newh * scale
@@ -233,7 +235,7 @@ class ViltFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestC
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)
feature_extractor_2 = self.feature_extraction_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)
for image in image_inputs:

27
tests/models/vit/test_feature_extraction_vit.py
View File

@@ -43,11 +43,12 @@ class ViTFeatureExtractionTester(unittest.TestCase):
min_resolution=30,
max_resolution=400,
do_resize=True,
size=18,
size=None,
do_normalize=True,
image_mean=[0.5, 0.5, 0.5],
image_std=[0.5, 0.5, 0.5],
):
size = size if size is not None else {"height": 18, "width": 18}
self.parent = parent
self.batch_size = batch_size
self.num_channels = num_channels
@@ -109,8 +110,8 @@ class ViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -121,8 +122,8 @@ class ViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -141,8 +142,8 @@ class ViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -153,8 +154,8 @@ class ViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -173,8 +174,8 @@ class ViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
1,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)
@@ -185,7 +186,7 @@ class ViTFeatureExtractionTest(FeatureExtractionSavingTestMixin, unittest.TestCa
(
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
self.feature_extract_tester.size,
self.feature_extract_tester.size,
self.feature_extract_tester.size["height"],
self.feature_extract_tester.size["width"],
),
)

71
tests/utils/test_image_processing_utils.py Normal file
View File

@@ -0,0 +1,71 @@
# coding=utf-8
# Copyright 2022 HuggingFace Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
from transformers.image_processing_utils import get_size_dict
class ImageProcessingUtilsTester(unittest.TestCase):
def test_get_size_dict(self):
# Test a dict with the wrong keys raises an error
inputs = {"wrong_key": 224}
with self.assertRaises(ValueError):
get_size_dict(inputs)
inputs = {"height": 224}
with self.assertRaises(ValueError):
get_size_dict(inputs)
inputs = {"width": 224, "shortest_edge": 224}
with self.assertRaises(ValueError):
get_size_dict(inputs)
# Test a dict with the correct keys is returned as is
inputs = {"height": 224, "width": 224}
outputs = get_size_dict(inputs)
self.assertEqual(outputs, inputs)
inputs = {"shortest_edge": 224}
outputs = get_size_dict(inputs)
self.assertEqual(outputs, {"shortest_edge": 224})
inputs = {"longest_edge": 224, "shortest_edge": 224}
outputs = get_size_dict(inputs)
self.assertEqual(outputs, {"longest_edge": 224, "shortest_edge": 224})
# Test a single int value which represents (size, size)
outputs = get_size_dict(224)
self.assertEqual(outputs, {"height": 224, "width": 224})
# Test a single int value which represents the shortest edge
outputs = get_size_dict(224, default_to_square=False)
self.assertEqual(outputs, {"shortest_edge": 224})
# Test a tuple of ints which represents (height, width)
outputs = get_size_dict((150, 200))
self.assertEqual(outputs, {"height": 150, "width": 200})
# Test a tuple of ints which represents (width, height)
outputs = get_size_dict((150, 200), height_width_order=False)
self.assertEqual(outputs, {"height": 200, "width": 150})
# Test an int representing the shortest edge and max_size which represents the longest edge
outputs = get_size_dict(224, max_size=256, default_to_square=False)
self.assertEqual(outputs, {"shortest_edge": 224, "longest_edge": 256})
# Test int with default_to_square=True and max_size fails
with self.assertRaises(ValueError):
get_size_dict(224, max_size=256, default_to_square=True)