Skip to content

set OMP_NUM_THREADS during clustering to ensure reproducibility #3400

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
KennethEnevoldsen merged 1 commit into from
Oct 19, 2025
Merged

set OMP_NUM_THREADS during clustering to ensure reproducibility #3400

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
13 changes: 13 additions & 0 deletions mteb/abstasks/clustering.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
import itertools
import logging
import os
import random
from collections import defaultdict
from pathlib import Path
Expand Down Expand Up @@ -31,6 +32,7 @@


MultilingualDataset = dict[HFSubset, DatasetDict]
OMP_NUM_THREADS = 4


def _evaluate_clustering_bootstrapped(
Expand All @@ -53,6 +55,17 @@ def _evaluate_clustering_bootstrapped(
- A dictionary where keys are level names (e.g., "Level 0", "Level 1", etc.) and values are lists of V-measure scores for each clustering experiment at that level.
- A dictionary where keys are level names and values are lists of cluster assignments for each clustering experiment at that level.
"""
# set OMP_NUM_THREADS for reproductibility
Copy link
Copy Markdown
Collaborator

@isaac-chung isaac-chung Oct 18, 2025

Choose a reason for hiding this comment

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

Can we add a sentence to describe how this ensures reproducibility please?

Copy link
Copy Markdown
Contributor Author

@KennethEnevoldsen KennethEnevoldsen Oct 19, 2025

Choose a reason for hiding this comment

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

Yeah, but I am not really sure - it was just the only thing that I expected would vary across CI, @Samoed, and my implementation. However, given that the tests are flaky, it is really hard to tell if it actually ensure anything.

Copy link
Copy Markdown
Contributor Author

@KennethEnevoldsen KennethEnevoldsen Oct 19, 2025

Choose a reason for hiding this comment

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

I will merge this for now, just to see if it resolves some of the flakiness - if it does not, I will roll it back.

if "OMP_NUM_THREADS" not in os.environ:
logger.info(
f"Setting OMP_NUM_THREADS to {OMP_NUM_THREADS} for clustering to ensure reproducibility."
)
os.environ["OMP_NUM_THREADS"] = str(OMP_NUM_THREADS)
else:
logger.info(
f"Using existing OMP_NUM_THREADS={os.environ['OMP_NUM_THREADS']} for clustering, this may lead to non-reproducible results. Set it to {OMP_NUM_THREADS} to ensure reproducibility."
)

v_measures = defaultdict(list)
cluster_assignments = defaultdict(list)
if max_depth is not None:
Expand Down