Skip to content

Fix resharding bug when dataset variable is reassigned #49

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
m-laprise wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix resharding bug when dataset variable is reassigned #49

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
5 changes: 3 additions & 2 deletions flame/data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -557,8 +557,9 @@ def build_dataset(
color = utils.Color
min_num_shards = dp_degree * num_workers if dp_degree else None
if len(dataset.split(',')) == 1:
dataset_path = dataset
Copy link

@gemini-code-assist gemini-code-assist bot Jan 14, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

Introducing dataset_path correctly fixes the bug. On a related note for maintainability, the load_dataset call on the next line is duplicated in the resharding logic (lines 588-597). Consider refactoring these calls to reduce duplication, for example by using a shared dictionary of arguments.

dataset = load_dataset(
path=dataset,
path=dataset_path,
name=dataset_name,
split=dataset_split,
data_dir=data_dir,
Expand All @@ -585,7 +586,7 @@ def build_dataset(
f"{color.reset}"
)
dataset = load_dataset(
path=dataset,
path=dataset_path,
name=dataset_name,
split=dataset_split,
data_dir=data_dir,
Expand Down