| Name | CVE-2025-46153 |
| Description | PyTorch 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. |
| Source | CVE (at NVD; CERT, ENISA, LWN, oss-sec, fulldisc, Debian ELTS, Red Hat, Ubuntu, Gentoo, SUSE bugzilla/CVE, GitHub advisories/code/issues, web search, more) |
| Debian Bugs | 1116538 |
The table below lists information on source packages.
| Source Package | Release | Version | Status |
|---|---|---|---|
| pytorch (PTS) | bullseye | 1.7.1-7 | vulnerable |
| bookworm | 1.13.1+dfsg-4 | vulnerable | |
| trixie | 2.6.0+dfsg-7 | vulnerable | |
| forky, sid | 2.6.0+dfsg-9 | vulnerable |
The information below is based on the following data on fixed versions.
| Package | Type | Release | Fixed Version | Urgency | Origin | Debian Bugs |
|---|---|---|---|---|---|---|
| pytorch | source | (unstable) | (unfixed) | 1116538 |
[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)