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 LayerNorm have a bias parameter attribute but is not instance of torch primitive modules #1229

Merged
merged 2 commits into from
Jun 28, 2023
Merged

Fix LayerNorm have a bias parameter attribute but is not instance of torch primitive modules #1229

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f48327e
Include additional norm classes into list of types that have bias
BloodAxe Jun 28, 2023
8792bca
Merge branch 'master' into feature/SG-841
BloodAxe Jun 28, 2023
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 src/super_gradients/training/utils/optimizer_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,11 +62,11 @@ def _get_no_decay_param_ids(module: nn.Module):
NOTE - ALL MODULES WITH ATTRIBUTES NAMED BIAS AND ARE INSTANCE OF nn.Parameter WILL BE CONSIDERED A BIAS PARAM FOR
ZERO WEIGHT DECAY.
"""
batchnorm_types = (_BatchNorm,)
norm_types = (_BatchNorm, nn.GroupNorm, nn.LayerNorm, nn.InstanceNorm1d, nn.InstanceNorm2d, nn.InstanceNorm3d)
torch_weight_with_bias_types = (_ConvNd, nn.Linear)
no_decay_ids = []
for name, m in module.named_modules():
if isinstance(m, batchnorm_types):
if isinstance(m, norm_types):
no_decay_ids.append(id(m.weight))
no_decay_ids.append(id(m.bias))
elif hasattr(m, "bias") and isinstance(m.bias, nn.Parameter):
Expand Down