CVE-2025-46153

NameCVE-2025-46153
DescriptionPyTorch before 3.7.0 has a bernoulli_p decompose function in decompositions.py even though it lacks full consistency with the eager CPU implementation, negatively affecting nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d for fallback_random=True.
SourceCVE (at NVD; CERT, LWN, oss-sec, fulldisc, Debian ELTS, Red Hat, Ubuntu, Gentoo, SUSE bugzilla/CVE, GitHub advisories/code/issues, web search, more)
Debian Bugs1116538

Vulnerable and fixed packages

The table below lists information on source packages.

Source PackageReleaseVersionStatus
pytorch (PTS)bullseye1.7.1-7vulnerable
bookworm1.13.1+dfsg-4vulnerable
trixie2.6.0+dfsg-7vulnerable
forky, sid2.6.0+dfsg-9vulnerable

The information below is based on the following data on fixed versions.

PackageTypeReleaseFixed VersionUrgencyOriginDebian Bugs
pytorchsource(unstable)(unfixed)1116538

Notes

[trixie] - pytorch <no-dsa> (Minor issue)
[bookworm] - pytorch <no-dsa> (Minor issue)
https://github.com/pytorch/pytorch/issues/142853
https://github.com/pytorch/pytorch/pull/143460
Fixed by: https://github.com/pytorch/pytorch/commit/288aa873831057b1eb7d747914ec4fdc76c23a80 (v2.7.0-rc1)

Search for package or bug name: Reporting problems