Skip to content

fix(scripts): single-device sharding for compute_norm_stats torch path #999

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
Bartok9 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix(scripts): single-device sharding for compute_norm_stats torch path #999

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
4 changes: 4 additions & 0 deletions scripts/compute_norm_stats.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
import openpi.training.config as _config
import openpi.training.data_loader as _data_loader
import openpi.transforms as transforms
import jax


class RemoveStrings(transforms.DataTransformFn):
Expand Down Expand Up @@ -47,9 +48,12 @@ def create_torch_dataloader(
else:
num_batches = len(dataset) // batch_size
shuffle = False
# Single-device sharding: batch size is often 1 for stats; multi-GPU default
# mesh sharding cannot split size-1 batches (see openpi#438).
data_loader = _data_loader.TorchDataLoader(
dataset,
local_batch_size=batch_size,
sharding=jax.sharding.SingleDeviceSharding(jax.devices()[0]),
num_workers=num_workers,
shuffle=shuffle,
num_batches=num_batches,
Expand Down