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

fix: Fixed load_pretrained_params in PyTorch when ignoring keys #902

Merged
merged 11 commits into from
Apr 28, 2022
Merged

fix: Fixed load_pretrained_params in PyTorch when ignoring keys #902

Changes from 1 commit
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
43814e3
fix: Fixed load_pretrained_params
Apr 27, 2022
f1f63a9
fix: Fixed ignore_keys for models in PyTorch
Apr 27, 2022
b51b217
test: Updated unittest accordingly
Apr 27, 2022
1bdedd2
refactor: Improved checks
Apr 27, 2022
2d28bf1
test: Added unittest for ignore_keys
Apr 27, 2022
b2a21dc
fix: Added missing import
Apr 28, 2022
79173f4
style: Fixed typing
Apr 28, 2022
de0652b
refactor: Refactored recognition ignore_keys
Apr 28, 2022
f8202ee
fix: Fixed typo
Apr 28, 2022
7de2ddc
fix: Fixed ResNet
Apr 28, 2022
5165d30
fix: Fixed load conditions
Apr 28, 2022
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
Prev Previous commit
Next Next commit
fix: Fixed ResNet
  • Loading branch information
FG Fernandez committed Apr 28, 2022
commit 7de2ddccf78e7b817292fdcfa967e2a747b286c5
11 changes: 4 additions & 7 deletions doctr/models/classification/resnet/pytorch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -175,13 +175,10 @@ def _resnet(
model = ResNet(num_blocks, output_channels, stage_stride, stage_conv, stage_pooling, cfg=_cfg, **kwargs)
# Load pretrained parameters
if pretrained:
_kwargs = {}
if kwargs['num_classes'] != len(default_cfgs[arch]['classes']):
# The number of classes is not the same as the number of classes in the pretrained model =>
# remove the last layer weights
_kwargs = {"ignore_keys": ignore_keys}
_kwargs = {"ignore_keys": ['13.weight', '13.bias']}
load_pretrained_params(model, default_cfgs[arch]['url'], **_kwargs)
# The number of classes is not the same as the number of classes in the pretrained model =>
# remove the last layer weights
_ignore_keys = ignore_keys if kwargs['num_classes'] == len(default_cfgs[arch]['classes']) else None
load_pretrained_params(model, default_cfgs[arch]['url'], ignore_keys=_ignore_keys)

return model

Expand Down