AI and data stack advisories

Severe, 6 weeks1496Projects326

1496 severe, 6 weeks · 326 projects

PyTorchPYSEC-2025-200

In PyTorch before 2.7.0, when torch.compile is used, FractionalMaxPool2d has inconsistent results

PyTorch

CVE-2025-46150 · Published Sep 25, 2025 · updated May 20, 2026

Medium5.3
Fix: upgrade to 2.7.0 or later
Source advisory

In PyTorch before 2.7.0, when torch.compile is used, FractionalMaxPool2d has inconsistent results.

Affected versions

PackageAffectedFixed in
torch
PyPI
>= 2.6.0, < 2.7.02.7.0
Details and references
CVSS 3.1
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N
Severity from
the CVSS score
Also known as
BIT-pytorch-2025-46150, CVE-2025-46150

More PyTorch advisories

All PyTorch
High 7.5Sep 25, 2025PyTorch: denial of serviceCVE-2025-55560fixed in 2.7.1High 7.5Sep 25, 2025pytorch v2.8.0 was discovered to display unexpected behaviorCVE-2025-55552fixed in 2.9.0High 7.5Sep 25, 2025PyTorch: denial of serviceCVE-2025-55553fixed in 2.7.1Medium 5.3Sep 25, 2025pytorch v2.8.0 was discovered to contain an integer overflow in the component torch.nan_to_num-.long()CVE-2025-55554fixed in 2.9.0High 7.5Sep 25, 2025PyTorch: denial of serviceCVE-2025-55557fixed in 2.7.1High 7.5Sep 25, 2025PyTorch: buffer overflowCVE-2025-55558fixed in 2.7.1