Vulnerability in PyTorch Affects Dropout Functions

CVE-2025-46153

What is CVE-2025-46153?

A specific issue has been identified in PyTorch where the bernoulli_p decompose function defined in decompositions.py does not achieve full consistency with the eager CPU implementation. This inconsistency can adversely affect the performance and reliability of nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d when the fallback_random option is enabled, potentially leading to unintended behavior in machine learning models relying on these dropout functionalities.

Human OS v1.0:
Ageing Is an Unpatched Zero-Day Vulnerability.

Remediate biological technical debt. Prime Ageing uses 95% high-purity SIRT6 activation to maintain genomic integrity and bolster systemic resilience.

Deploy the Patch →

References