Skip to content

fix: respect gradient checkpointing configuration #1

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

Open
XieWeikai wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: respect gradient checkpointing configuration #1

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
7 changes: 0 additions & 7 deletions src/openpi/models_pytorch/gemma_pytorch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,13 +134,6 @@ def forward(
and self.training
) or (hasattr(self, "gradient_checkpointing") and self.gradient_checkpointing and self.training)

# Force enable gradient checkpointing if we're in training mode and the model supports it
if self.training and hasattr(self.gemma_expert.model, "gradient_checkpointing"):
if not self.gemma_expert.model.gradient_checkpointing:
print("Forcing gradient checkpointing to be enabled for Gemma expert model")
self.gemma_expert.model.gradient_checkpointing = True
use_gradient_checkpointing = True

# Debug gradient checkpointing status
if hasattr(self, "_debug_gc_printed") and not self._debug_gc_printed:
print(f"Gemma expert model gradient checkpointing: {use_gradient_checkpointing}")
Expand Down