Skip to content

Revert "Type cast before normalize maskformer" #51

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
amyeroberts merged 1 commit into from
Sep 21, 2022
Merged

Revert "Type cast before normalize maskformer" #51

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 4 additions & 17 deletions src/transformers/models/maskformer/feature_extraction_maskformer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -206,11 +206,9 @@ def __call__(
instance id. To convert it to a binary mask of shape (`batch, num_labels, height, width`) we need a
dictionary mapping instance ids to label ids to create a semantic segmentation map.

return_tensors (`str` or [`~utils.TensorType`], *optional*, defaults to `None`):
If set, will return a tensor of a particular framework.

Acceptable values are:
- `'pt'`: Return PyTorch `torch.Tensor` objects.
return_tensors (`str` or [`~file_utils.TensorType`], *optional*):
If set, will return tensors instead of NumPy arrays. If set to `'pt'`, return PyTorch `torch.Tensor`
objects.

Returns:
[`BatchFeature`]: A [`BatchFeature`] with the following fields:
Expand Down Expand Up @@ -285,19 +283,8 @@ def __call__(
image=image, target=None, size=self.size, max_size=self.max_size
)[0]

# if do_normalize=False, the casting to a numpy array won't happen, so we need to do it here
make_channel_first = True if isinstance(images[0], Image.Image) else images[0].shape[-1] in (1, 3)
images = [self.to_numpy_array(image, rescale=False, channel_first=make_channel_first) for image in images]
if segmentation_maps is not None:
segmentation_maps = [
self.to_numpy_array(segmap, rescale=False, channel_first=True) for segmap in segmentation_maps
]

if self.do_normalize:
images = [
self.normalize(image=image, mean=self.image_mean, std=self.image_std, rescale=True) for image in images
]

images = [self.normalize(image=image, mean=self.image_mean, std=self.image_std) for image in images]
# NOTE I will be always forced to pad them them since they have to be stacked in the batch dim
encoded_inputs = self.encode_inputs(
images,
Expand Down
41 changes: 0 additions & 41 deletions tests/models/maskformer/test_feature_extraction_maskformer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,6 @@

import numpy as np

from parameterized import parameterized
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available

Expand Down Expand Up @@ -402,43 +401,3 @@ def test_post_process_panoptic_segmentation(self):
self.assertEqual(
el["segmentation"].shape, (self.feature_extract_tester.height, self.feature_extract_tester.width)
)

@require_torch
@parameterized.expand(
[
("do_resize_True_do_normalize_True", True, True),
("do_resize_True_do_normalize_False", True, False),
("do_resize_True_do_normalize_True", True, True),
("do_resize_True_do_normalize_False", True, False),
("do_resize_False_do_normalize_True", False, True),
("do_resize_False_do_normalize_False", False, False),
("do_resize_False_do_normalize_True", False, True),
("do_resize_False_do_normalize_False", False, False),
]
)
def test_call_flags(self, _, do_resize, do_normalize):
# Initialize feature_extractor
feature_extractor = self.feature_extraction_class(**self.feat_extract_dict)
feature_extractor.do_resize = do_resize
feature_extractor.do_normalize = do_normalize
# create random PIL images
image_inputs = prepare_image_inputs(self.feature_extract_tester, equal_resolution=False)

all_image_shapes = [img.size[::-1] for img in image_inputs]
if do_resize:
all_image_shapes = [
self.feature_extract_tester.get_expected_values([image], batched=False) for image in image_inputs
]

max_across_dim = [max(shape) for shape in zip(*all_image_shapes)]
expected_shape = (
self.feature_extract_tester.batch_size,
self.feature_extract_tester.num_channels,
*max_across_dim,
)

pixel_values = feature_extractor(image_inputs, return_tensors="pt")["pixel_values"]
self.assertEqual(len(pixel_values), self.feature_extract_tester.batch_size)

self.assertEqual(pixel_values.shape, expected_shape)
self.assertIsInstance(pixel_values, torch.Tensor)