Bug#1116538: pytorch: CVE-2025-46153

[Salvatore Bonaccorso <carnil@debian.org>]

Source: pytorch
Version: 2.6.0+dfsg-9
Severity: important
Tags: security upstream
Forwarded: https://github.com/pytorch/pytorch/issues/142853
X-Debbugs-Cc: carnil@debian.org, Debian Security Team <team@security.debian.org>

Hi,

The following vulnerability was published for pytorch.

CVE-2025-46153[0]:
| 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.


If you fix the vulnerability please also make sure to include the
CVE (Common Vulnerabilities & Exposures) id in your changelog entry.

For further information see:

[0] https://security-tracker.debian.org/tracker/CVE-2025-46153
    https://www.cve.org/CVERecord?id=CVE-2025-46153
[1] https://github.com/pytorch/pytorch/issues/142853
[2] https://github.com/pytorch/pytorch/pull/143460
[3] https://github.com/pytorch/pytorch/commit/288aa873831057b1eb7d747914ec4fdc76c23a80

Please adjust the affected versions in the BTS as needed.

Regards,
Salvatore