Skip to content

Profile Trainer.save_checkpoint #20405

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
merged 4 commits into from
Nov 13, 2024
Merged

Profile Trainer.save_checkpoint #20405

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
d17e4e5
Profile Trainer.save_checkpoint
EricCousineau-TRI Nov 7, 2024
2aece26
Merge branch 'master' into feature-save-ckpt
lantiga Nov 13, 2024
a4b0095
Merge branch 'master' into feature-save-ckpt
t-vi Nov 13, 2024
fcc95af
Merge branch 'master' into feature-save-ckpt
Borda Nov 13, 2024
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: 4 additions & 3 deletions src/lightning/pytorch/trainer/trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1361,9 +1361,10 @@ def save_checkpoint(
"Saving a checkpoint is only possible if a model is attached to the Trainer. Did you call"
" `Trainer.save_checkpoint()` before calling `Trainer.{fit,validate,test,predict}`?"
)
checkpoint = self._checkpoint_connector.dump_checkpoint(weights_only)
self.strategy.save_checkpoint(checkpoint, filepath, storage_options=storage_options)
self.strategy.barrier("Trainer.save_checkpoint")
with self.profiler.profile("save_checkpoint"):
checkpoint = self._checkpoint_connector.dump_checkpoint(weights_only)
self.strategy.save_checkpoint(checkpoint, filepath, storage_options=storage_options)
self.strategy.barrier("Trainer.save_checkpoint")

"""
State properties
Expand Down
Loading