Skip to content

ReturnnTrainingJob, support JAX checkpoints #662

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

ReturnnTrainingJob, support JAX checkpoints #662

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: 7 additions & 0 deletions returnn/training.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -218,6 +218,13 @@ def __init__(
for pt_path in [self.output_path("models/epoch.%.3d.pt" % k)]
}
self.out_models = None
elif self.returnn_config.get("backend", None) == "jax":
# A directory, not a file.
# Not directory=True: pre-creating them empty could confuse other code.
self.out_checkpoints = {
k: self.output_path("models/epoch.%.3d.orbax" % k) for k in stored_epochs if k in self.keep_epochs
}
self.out_models = None
else:
raise ValueError("'backend' not specified in config")

Expand Down
Loading