Skip to content

Updated _load_pretrained_model_low_mem to check if keys are in the state_dict #16643

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
FrancescoSaverioZuppichini merged 7 commits into from
Apr 7, 2022
Merged

Updated _load_pretrained_model_low_mem to check if keys are in the state_dict #16643

Changes from 4 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
089964f
Updated _load_pretrained_model_low_mem to check if keys are in the st…
FrancescoSaverioZuppichini Apr 7, 2022
db383cc
update after conversions
FrancescoSaverioZuppichini Apr 7, 2022
136f46f
update after conversation
FrancescoSaverioZuppichini Apr 7, 2022
30e715b
update after conversation
FrancescoSaverioZuppichini Apr 7, 2022
cef4d7f
update after conversation
FrancescoSaverioZuppichini Apr 7, 2022
5f71190
update after conversation (I should sleep)
FrancescoSaverioZuppichini Apr 7, 2022
10131af
update after conversation
FrancescoSaverioZuppichini Apr 7, 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
15 changes: 8 additions & 7 deletions src/transformers/modeling_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2162,13 +2162,14 @@ def find_submodule_and_param_name(model, long_key):

# materialize state_dict entries one by one on CPU
for k in loaded_state_dict_keys:
submodule, param_name = find_submodule_and_param_name(model, k)
if submodule is not None:
param_dtype = getattr(submodule, param_name).dtype
new_val = state_dict[k].to(param_dtype)
if isinstance(getattr(submodule, param_name), torch.nn.Parameter):
new_val = torch.nn.Parameter(new_val)
setattr(submodule, param_name, new_val)
if k in state_dict:
submodule, param_name = find_submodule_and_param_name(model, k)
if submodule is not None:
param_dtype = getattr(submodule, param_name).dtype
if isinstance(getattr(submodule, param_name), torch.nn.Parameter):
new_val = torch.nn.Parameter(new_val)
setattr(submodule, param_name, new_val)
new_val = state_dict[k].to(param_dtype)

del state_dict

Expand Down