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

Del ort_model._modules to foward its accessing to torch_model._modules #14563

Merged
merged 1 commit into from
Mar 3, 2023
Merged

Del ort_model._modules to foward its accessing to torch_model._modules #14563

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
4 changes: 4 additions & 0 deletions orttraining/orttraining/python/training/ortmodule/ortmodule.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,6 +104,10 @@ def __init__(self, module, debug_options=None):
# else, they will be assigned to self._torch_module.original_module instead.
self._is_initialized = True

# del the ort._modules so that all reference to ort._modules will be forward to the underlying torch_model
# through '__getattr__'
del self._modules

# IMPORTANT: DO NOT add code here
# This declaration is for automatic document generation purposes only
# The actual forward implementation is bound during ORTModule initialization
Expand Down
3 changes: 2 additions & 1 deletion orttraining/orttraining/test/python/orttraining_test_ortmodule_api.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4192,7 +4192,8 @@ def forward(self, x):
x = torch.randn(N, D_in, device=device)
_ = wrapper_module(x)

state_dict1 = wrapper_module.state_dict()
# Must copy the state_dict or else they are sharing the same memory
state_dict1 = copy.deepcopy(wrapper_module.state_dict())
list(next(iter(state_dict1.items())))[1] += 10
wrapper_module.load_state_dict(state_dict1)
state_dict2 = wrapper_module.state_dict()
Expand Down