CVE-2025-46153

Published: Sep 25, 2025Modified: Jun 17, 2026

5.3

Vector

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N

Exploitability: 3.9 / Impact: 1.4

Source: 134c704f-9b21-4f2e-91b3-4a467353bcc0 (Secondary)

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.

Affected (1)

Products: Linuxfoundation: Pytorch

Linuxfoundation

1 product

Pytorch

Configuration A

1 vulnerable

Vulnerable Software	Affected Versions
Linuxfoundation Pytorch	From 2.6.0 to 2.7.0

Related CWEs

CWE-1176

Inefficient CPU Computation

The product performs CPU computations using algorithms that are not as efficient as they could be for the needs of the developer, i.e., the computations can be optimized further.

References (5)

https://gist.github.com/shaoyuyoung/4bcefba4004f8271e64b5185c95a248a

Source: cve@mitre.org

Third Party Advisory

https://gist.github.com/shaoyuyoung/e636f2e7a306105b7e96809e2b85c28a

Source: cve@mitre.org

Third Party Advisory

https://github.com/pytorch/pytorch/compare/v2.6.0...v2.7.0

Source: cve@mitre.org

Product

https://github.com/pytorch/pytorch/issues/142853

Source: cve@mitre.org

Issue Tracking

https://github.com/pytorch/pytorch/pull/143460

Source: cve@mitre.org

Issue TrackingPatch

Timeline

No history available yet.