Refactor image processor testers (#25450)
* Refactor image processor test mixin - Move test_call_numpy, test_call_pytorch, test_call_pil to mixin - Rename mixin to reflect handling of logic more than saving - Add prepare_image_inputs, expected_image_outputs for tests * Fix for oneformer
This commit is contained in:
amyeroberts
@@ -16,20 +16,13 @@
|
||||
|
||||
import unittest
|
||||
|
||||
import numpy as np
|
||||
|
||||
from transformers.testing_utils import require_torch, require_vision
|
||||
from transformers.utils import is_torch_available, is_vision_available
|
||||
from transformers.utils import is_vision_available
|
||||
|
||||
from ...test_image_processing_common import ImageProcessingSavingTestMixin
|
||||
from ...test_image_processing_common import ImageProcessingTestMixin, prepare_image_inputs
|
||||
|
||||
|
||||
if is_torch_available():
|
||||
import torch
|
||||
|
||||
if is_vision_available():
|
||||
from PIL import Image
|
||||
|
||||
from transformers import BlipImageProcessor
|
||||
|
||||
|
||||
@@ -76,40 +69,24 @@ class BlipImageProcessingTester(unittest.TestCase):
|
||||
"do_pad": self.do_pad,
|
||||
}
|
||||
|
||||
def prepare_inputs(self, equal_resolution=False, numpify=False, torchify=False):
|
||||
"""This function prepares a list of PIL images, or a list of numpy arrays if one specifies numpify=True,
|
||||
or a list of PyTorch tensors if one specifies torchify=True.
|
||||
"""
|
||||
def expected_output_image_shape(self, images):
|
||||
return 3, self.size["height"], self.size["width"]
|
||||
|
||||
assert not (numpify and torchify), "You cannot specify both numpy and PyTorch tensors at the same time"
|
||||
|
||||
if equal_resolution:
|
||||
image_inputs = []
|
||||
for i in range(self.batch_size):
|
||||
image_inputs.append(
|
||||
np.random.randint(
|
||||
255, size=(self.num_channels, self.max_resolution, self.max_resolution), dtype=np.uint8
|
||||
)
|
||||
)
|
||||
else:
|
||||
image_inputs = []
|
||||
for i in range(self.batch_size):
|
||||
width, height = np.random.choice(np.arange(self.min_resolution, self.max_resolution), 2)
|
||||
image_inputs.append(np.random.randint(255, size=(self.num_channels, width, height), dtype=np.uint8))
|
||||
|
||||
if not numpify and not torchify:
|
||||
# PIL expects the channel dimension as last dimension
|
||||
image_inputs = [Image.fromarray(np.moveaxis(x, 0, -1)) for x in image_inputs]
|
||||
|
||||
if torchify:
|
||||
image_inputs = [torch.from_numpy(x) for x in image_inputs]
|
||||
|
||||
return image_inputs
|
||||
def prepare_image_inputs(self, equal_resolution=False, numpify=False, torchify=False):
|
||||
return prepare_image_inputs(
|
||||
batch_size=self.batch_size,
|
||||
num_channels=self.num_channels,
|
||||
min_resolution=self.min_resolution,
|
||||
max_resolution=self.max_resolution,
|
||||
equal_resolution=equal_resolution,
|
||||
numpify=numpify,
|
||||
torchify=torchify,
|
||||
)
|
||||
|
||||
|
||||
@require_torch
|
||||
@require_vision
|
||||
class BlipImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase):
|
||||
class BlipImageProcessingTest(ImageProcessingTestMixin, unittest.TestCase):
|
||||
image_processing_class = BlipImageProcessor if is_vision_available() else None
|
||||
|
||||
def setUp(self):
|
||||
@@ -128,109 +105,10 @@ class BlipImageProcessingTest(ImageProcessingSavingTestMixin, unittest.TestCase)
|
||||
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 image_processor
|
||||
image_processor = self.image_processing_class(**self.image_processor_dict)
|
||||
# create random PIL images
|
||||
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 = image_processor(image_inputs[0], return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
1,
|
||||
self.image_processor_tester.num_channels,
|
||||
self.image_processor_tester.size["height"],
|
||||
self.image_processor_tester.size["width"],
|
||||
),
|
||||
)
|
||||
|
||||
# Test batched
|
||||
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
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 image_processor
|
||||
image_processor = self.image_processing_class(**self.image_processor_dict)
|
||||
# create random numpy tensors
|
||||
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 = image_processor(image_inputs[0], return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
1,
|
||||
self.image_processor_tester.num_channels,
|
||||
self.image_processor_tester.size["height"],
|
||||
self.image_processor_tester.size["width"],
|
||||
),
|
||||
)
|
||||
|
||||
# Test batched
|
||||
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
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 image_processor
|
||||
image_processor = self.image_processing_class(**self.image_processor_dict)
|
||||
# create random PyTorch tensors
|
||||
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 = image_processor(image_inputs[0], return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
1,
|
||||
self.image_processor_tester.num_channels,
|
||||
self.image_processor_tester.size["height"],
|
||||
self.image_processor_tester.size["width"],
|
||||
),
|
||||
)
|
||||
|
||||
# Test batched
|
||||
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
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(ImageProcessingSavingTestMixin, unittest.TestCase):
|
||||
class BlipImageProcessingTestFourChannels(ImageProcessingTestMixin, unittest.TestCase):
|
||||
image_processing_class = BlipImageProcessor if is_vision_available() else None
|
||||
|
||||
def setUp(self):
|
||||
@@ -250,37 +128,10 @@ class BlipImageProcessingTestFourChannels(ImageProcessingSavingTestMixin, unitte
|
||||
self.assertTrue(hasattr(image_processor, "image_std"))
|
||||
self.assertTrue(hasattr(image_processor, "do_convert_rgb"))
|
||||
|
||||
def test_batch_feature(self):
|
||||
pass
|
||||
@unittest.skip("BlipImageProcessor does not support 4 channels yet") # FIXME Amy
|
||||
def test_call_numpy(self):
|
||||
return super().test_call_numpy()
|
||||
|
||||
def test_call_pil_four_channels(self):
|
||||
# Initialize image_processor
|
||||
image_processor = self.image_processing_class(**self.image_processor_dict)
|
||||
# create random PIL images
|
||||
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 = image_processor(image_inputs[0], return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
1,
|
||||
self.expected_encoded_image_num_channels,
|
||||
self.image_processor_tester.size["height"],
|
||||
self.image_processor_tester.size["width"],
|
||||
),
|
||||
)
|
||||
|
||||
# Test batched
|
||||
encoded_images = image_processor(image_inputs, return_tensors="pt").pixel_values
|
||||
self.assertEqual(
|
||||
encoded_images.shape,
|
||||
(
|
||||
self.image_processor_tester.batch_size,
|
||||
self.expected_encoded_image_num_channels,
|
||||
self.image_processor_tester.size["height"],
|
||||
self.image_processor_tester.size["width"],
|
||||
),
|
||||
)
|
||||
@unittest.skip("BlipImageProcessor does not support 4 channels yet") # FIXME Amy
|
||||
def test_call_pytorch(self):
|
||||
return super().test_call_torch()
|
||||
|
||||
Reference in New Issue
Block a user