CVE-2025-46153: n/a
PyTorch versions prior to 3.7.0 contain an inconsistency in the bernoulli_p decompose function within decompositions.py. This inconsistency affects the behavior of nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d modules when the fallback_random parameter is set to true. The issue results in a deviation from the expected eager CPU implementation behavior.
AI Analysis
Technical Summary
CVE-2025-46153 identifies a functional inconsistency in PyTorch before version 3.7.0 related to the bernoulli_p decompose function in decompositions.py. This inconsistency impacts the nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d layers when fallback_random is enabled, causing them to behave differently than the eager CPU implementation. The vulnerability is classified under CWE-1176 and has a CVSS 3.1 base score of 5.3, indicating a medium severity level. No known exploits are reported in the wild, and no official patch or remediation link is provided in the data.
Potential Impact
The inconsistency in the bernoulli_p decompose function may lead to unexpected or incorrect behavior in dropout layers (nn.Dropout1d, nn.Dropout2d, nn.Dropout3d) when fallback_random is true. The impact is limited to a loss of consistency in the function's behavior, with a confidentiality impact rated low and no integrity or availability impact reported.
Mitigation Recommendations
Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Since no official patch or fix is referenced, users should monitor PyTorch releases for updates addressing this issue. Until a fix is available, avoid using fallback_random=true with the affected dropout modules if consistent behavior is critical.
CVE-2025-46153: n/a
Description
PyTorch versions prior to 3.7.0 contain an inconsistency in the bernoulli_p decompose function within decompositions.py. This inconsistency affects the behavior of nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d modules when the fallback_random parameter is set to true. The issue results in a deviation from the expected eager CPU implementation behavior.
CVSS v3.1
Score 5.3medium
Affected software
Weaknesses
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
CVE-2025-46153 identifies a functional inconsistency in PyTorch before version 3.7.0 related to the bernoulli_p decompose function in decompositions.py. This inconsistency impacts the nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d layers when fallback_random is enabled, causing them to behave differently than the eager CPU implementation. The vulnerability is classified under CWE-1176 and has a CVSS 3.1 base score of 5.3, indicating a medium severity level. No known exploits are reported in the wild, and no official patch or remediation link is provided in the data.
Potential Impact
The inconsistency in the bernoulli_p decompose function may lead to unexpected or incorrect behavior in dropout layers (nn.Dropout1d, nn.Dropout2d, nn.Dropout3d) when fallback_random is true. The impact is limited to a loss of consistency in the function's behavior, with a confidentiality impact rated low and no integrity or availability impact reported.
Mitigation Recommendations
Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Since no official patch or fix is referenced, users should monitor PyTorch releases for updates addressing this issue. Until a fix is available, avoid using fallback_random=true with the affected dropout modules if consistent behavior is critical.
Technical Details
- Data Version
- 5.1
- Assigner Short Name
- mitre
- Date Reserved
- 2025-04-22T00:00:00.000Z
- State
- PUBLISHED
Threat ID: 68d5511823f14e593ee333a8
Added to database: 09/25/2025, 14:26:32 UTC
Last enriched: 09/08/2026, 13:22:32 UTC
Last updated: 09/10/2026, 19:36:50 UTC
Views: 211
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.