CVE-2025-46153

torch: PyTorch inconsistent results

Published Sep 25, 2025 CVSS 3.1 · 5.3 Medium EPSS < 1% KEV · Not listed
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.

Remediation

No solution or workaround provided in the CVE record

Tracking

Sign in to view the affected projects.

Advisories

Source ID Title
EUVD EUVD EUVD-2025-31108 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.
History
Type Values Removed Values Added
First Time appeared Linuxfoundation
Linuxfoundation pytorch
CPEs cpe:2.3:a:linuxfoundation:pytorch:*:-:*:*:*:python:*:*
Vendors & Products Linuxfoundation
Linuxfoundation pytorch
Type Values Removed Values Added
First Time appeared Pytorch
Pytorch pytorch
Vendors & Products Pytorch
Pytorch pytorch
Type Values Removed Values Added
Weaknesses CWE-1176
Metrics cvssV3_1

{'score': 5.3, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N'}

ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'none', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Type Values Removed Values Added
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.
References
Weaknesses

Sources

Status
PUBLISHED
Assigner
mitre
Published
Updated
Sep 25, 2025, 6:33 PM
Reserved
Apr 22, 2025, 12:00 AM
Status
Analyzed
Published
Sep 25, 2025, 3:16 PM
Modified
Nov 5, 2026, 11:10 PM
Severity
Moderate
Public date
Sep 25, 2025, 12:00 AM