Skip to content
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

Jump to bottom

[SegFormer] Add support for segmentation masks with one label #20279

Merged
merged 4 commits into from
Dec 20, 2022
Merged

[SegFormer] Add support for segmentation masks with one label #20279

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
4404b7f
Add support for binary segmentation
Nov 16, 2022
f2d36ec
Fix loss calculation and add test
Nov 16, 2022
1367ba3
Remove space
Nov 16, 2022
6e4b8e3
use fstring
Dec 20, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
19 changes: 12 additions & 7 deletions src/transformers/models/segformer/modeling_segformer.py
View file
Open in desktop
Expand Up @@ -806,15 +806,20 @@ def forward(

loss = None
if labels is not None:
if not self.config.num_labels > 1:
raise ValueError("The number of labels should be greater than one")
else:
# upsample logits to the images' original size
upsampled_logits = nn.functional.interpolate(
logits, size=labels.shape[-2:], mode="bilinear", align_corners=False
)
# upsample logits to the images' original size
upsampled_logits = nn.functional.interpolate(
logits, size=labels.shape[-2:], mode="bilinear", align_corners=False
)
if self.config.num_labels > 1:
loss_fct = CrossEntropyLoss(ignore_index=self.config.semantic_loss_ignore_index)
loss = loss_fct(upsampled_logits, labels)
elif self.config.num_labels == 1:
valid_mask = ((labels >= 0) & (labels != self.config.semantic_loss_ignore_index)).float()
loss_fct = BCEWithLogitsLoss(reduction="none")
NielsRogge marked this conversation as resolved.
Show resolved Hide resolved
loss = loss_fct(upsampled_logits.squeeze(1), labels.float())
loss = (loss * valid_mask).mean()
else:
raise ValueError(f"Number of labels should be >=0: {self.config.num_labels}")

if not return_dict:
if output_hidden_states:
Expand Down
14 changes: 14 additions & 0 deletions tests/models/segformer/test_modeling_segformer.py
View file
Open in desktop
Expand Up @@ -140,6 +140,16 @@ def create_and_check_for_image_segmentation(self, config, pixel_values, labels):
self.parent.assertEqual(
result.logits.shape, (self.batch_size, self.num_labels, self.image_size // 4, self.image_size // 4)
)
self.parent.assertGreater(result.loss, 0.0)

def create_and_check_for_binary_image_segmentation(self, config, pixel_values, labels):
config.num_labels = 1
model = SegformerForSemanticSegmentation(config=config)
model.to(torch_device)
model.eval()
labels = torch.randint(0, 1, (self.batch_size, self.image_size, self.image_size)).to(torch_device)
result = model(pixel_values, labels=labels)
self.parent.assertGreater(result.loss, 0.0)

def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
Expand Down Expand Up @@ -177,6 +187,10 @@ def test_model(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*config_and_inputs)

def test_for_binary_image_segmentation(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_binary_image_segmentation(*config_and_inputs)

def test_for_image_segmentation(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_segmentation(*config_and_inputs)
Expand Down