Skip to content

Fix incorrectly frozen BN on ResNet FPN backbone #3396

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
merged 3 commits into from
Feb 15, 2021
Merged

Fix incorrectly frozen BN on ResNet FPN backbone #3396

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
ed5961b
Avoid freezing bn1 if all layers are trainable.
datumbox Feb 15, 2021
900dbe2
Merge branch 'master' into bugfix/resnet_fpn_bn_freeze
datumbox Feb 15, 2021
a2644df
Remove misleading comments.
datumbox Feb 15, 2021
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
4 changes: 2 additions & 2 deletions torchvision/models/detection/backbone_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -96,7 +96,8 @@ def resnet_fpn_backbone(
# select layers that wont be frozen
assert 0 <= trainable_layers <= 5
layers_to_train = ['layer4', 'layer3', 'layer2', 'layer1', 'conv1'][:trainable_layers]
# freeze layers only if pretrained backbone is used
if trainable_layers == 5:
layers_to_train.append('bn1')
for name, parameter in backbone.named_parameters():
if all([not name.startswith(layer) for layer in layers_to_train]):
parameter.requires_grad_(False)
Expand Down Expand Up @@ -152,7 +153,6 @@ def mobilenet_backbone(
assert 0 <= trainable_layers <= num_stages
freeze_before = num_stages if trainable_layers == 0 else stage_indices[num_stages - trainable_layers]

# freeze layers only if pretrained backbone is used
for b in backbone[:freeze_before]:
for parameter in b.parameters():
parameter.requires_grad_(False)
Expand Down