Inconsistent Results in FractionalMaxPool2d in PyTorch Prior to 2.7.0
CVE-2025-46150

5.3MEDIUM

Key Information:

Vendor

PyTorch

Status
Vendor
CVE Published:
25 September 2025

What is CVE-2025-46150?

Prior to version 2.7.0, PyTorch exhibits sporadic inconsistency in the results produced by the FractionalMaxPool2d function when utilizing torch.compile. This variability can lead to discrepancies in model performance, impacting users' ability to rely on the accuracy and reliability of their neural network outputs. It's essential for developers and users adopting PyTorch to understand these inconsistencies to mitigate potential issues in their applications and ensure robust cybersecurity measures are in place.

References

CVSS V3.1

Score:
5.3
Severity:
MEDIUM
Confidentiality:
Low
Integrity:
None
Availability:
Low
Attack Vector:
Network
Attack Complexity:
Low
Privileges Required:
None
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.