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

Add support for gradient checkpointing in BERT #4659

Merged
merged 9 commits into from
Jun 22, 2020
Merged

Add support for gradient checkpointing in BERT #4659

Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
7a0adbb
add support for gradient checkpointing in BERT
ibeltagy May 29, 2020
1765a14
fix unit tests
ibeltagy May 29, 2020
bf43427
isort
ibeltagy May 29, 2020
5f20845
Merge branch 'master' into grad_checkpointing
ibeltagy Jun 11, 2020
b36648f
black
ibeltagy Jun 11, 2020
5eb68bb
workaround for `torch.utils.checkpoint.checkpoint` not accepting bool
ibeltagy Jun 12, 2020
afee616
Revert "workaround for `torch.utils.checkpoint.checkpoint` not accept…
ibeltagy Jun 12, 2020
00bd3d5
workaround for `torch.utils.checkpoint.checkpoint` not accepting bool
ibeltagy Jun 12, 2020
400070b
Merge branch 'master' into grad_checkpointing
LysandreJik Jun 22, 2020
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 unit tests
  • Loading branch information
@ibeltagy
ibeltagy committed May 29, 2020
commit 1765a1430d091e0464a2413354eac362628db56a
10 changes: 7 additions & 3 deletions src/transformers/modeling_bert.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -405,10 +405,14 @@ def forward(
if self.output_hidden_states:
all_hidden_states = all_hidden_states + (hidden_states,)

if self.config.gradient_checkpointing:
if getattr(self.config, "gradient_checkpointing", False):
layer_outputs = torch.utils.checkpoint.checkpoint(
layer_module, hidden_states, attention_mask, head_mask[i],
encoder_hidden_states, encoder_attention_mask
layer_module,
hidden_states,
attention_mask,
head_mask[i],
encoder_hidden_states,
encoder_attention_mask,
)
else:
layer_outputs = layer_module(
Expand Down