Skip to content

Fix memory issues when computing retrieval metrics with high-valued indices #3291

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
wants to merge 3 commits into
base: master
Choose a base branch
from
Open

Fix memory issues when computing retrieval metrics with high-valued indices #3291

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
0e4be7f
Fix memory issues when computing retrieval metrics with high-valued i…
adaliaramon Oct 16, 2025
7159f57
Merge branch 'master' into fix-retrieval-memory-issues
justusschock Oct 16, 2025
d1d0a12
Merge branch 'master' into fix-retrieval-memory-issues
justusschock Oct 16, 2025
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
9 changes: 2 additions & 7 deletions src/torchmetrics/utilities/data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -230,13 +230,8 @@ def _flexible_bincount(x: Tensor) -> Tensor:
Number of occurrences for each unique element in x

"""
# make sure elements in x start from 0
x = x - x.min()
unique_x = torch.unique(x)

output = _bincount(x, minlength=torch.max(unique_x) + 1) # type: ignore[arg-type]
# remove zeros from output tensor
return output[unique_x]
unique_x, inverse_indices = torch.unique(x, return_inverse=True)
return _bincount(inverse_indices, minlength=len(unique_x)) # type: ignore[arg-type]


def allclose(tensor1: Tensor, tensor2: Tensor) -> bool:
Expand Down
Loading