Is CVE-2025-2953 exploited in the wild?

No. Junglewise Threat Intelligence has no report of CVE-2025-2953 being exploited in the wild.

How severe is CVE-2025-2953?

Low severity, with a CVSS score of 3.3, published 30 March 2025.

Which products and versions are affected?

torch (PyPI) < 2.7.1-rc1, ECOSYSTEM:introduced=0,fixed=2.7.1-rc1, ECOSYSTEM:introduced=0,last_affected=2.6.0+cu124; PyTorch < 2.7.1-rc1.

Is there a fix for CVE-2025-2953?

Yes. torch (PyPI): upgrade to 2.7.1-rc1 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

How is CVE-2025-2953 attacked?

It is attacked with local access, in a software library.

Am I affected by CVE-2025-2953?

Check whether you run torch (PyPI) or PyTorch at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

CVE-2025-2953: PyTorch denial of service in torch.mkldnn_max_pool2d

Q: Which products and versions are affected?

torch (PyPI) < 2.7.1-rc1, ECOSYSTEM:introduced=0,fixed=2.7.1-rc1, ECOSYSTEM:introduced=0,last_affected=2.6.0+cu124; PyTorch < 2.7.1-rc1.

Q: Is there a fix for CVE-2025-2953?

Yes. torch (PyPI): upgrade to 2.7.1-rc1 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

Q: How is CVE-2025-2953 attacked?

It is attacked with local access, in a software library.

Q: Am I affected by CVE-2025-2953?

Check whether you run torch (PyPI) or PyTorch at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

CVE-2025-2953 · Severity: low · CVSS 3.3 · Published 2025-03-30

PyTorch, a popular machine learning framework, is vulnerable to a local denial-of-service attack. An attacker with local access to a system running PyTorch can trigger a crash by providing specific parameters to a pooling function. This could lead to the interruption of machine learning workloads or service outages on affected systems.

Technologies: torch (PyPI), PyTorch. Vendors: PyPI, PyTorch.

Quick answers

Is CVE-2025-2953 exploited in the wild?: No. Junglewise Threat Intelligence has no report of CVE-2025-2953 being exploited in the wild.
How severe is CVE-2025-2953?: Low severity, with a CVSS score of 3.3, published 30 March 2025.
Which products and versions are affected?: torch (PyPI) < 2.7.1-rc1, ECOSYSTEM:introduced=0,fixed=2.7.1-rc1, ECOSYSTEM:introduced=0,last_affected=2.6.0+cu124; PyTorch < 2.7.1-rc1.
Is there a fix for CVE-2025-2953?: Yes. torch (PyPI): upgrade to 2.7.1-rc1 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2025-2953 attacked?: It is attacked with local access, in a software library.
Am I affected by CVE-2025-2953?: Check whether you run torch (PyPI) or PyTorch at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

How to fix

torch (PyPI): upgrade to 2.7.1-rc1 or later
```
pip install "torch>=2.7.1-rc1"
```

Technical details

A vulnerability in PyTorch versions prior to 2.7.1-rc1 allows a local attacker to cause a denial of service (DoS) via a floating point exception. The issue resides in the 'torch.mkldnn_max_pool2d' function, specifically when handling certain input parameters such as a 'stride' value of zero. This lack of proper error checking in the MKLDNN (oneDNN) integration leads to a process crash. Exploitation requires the ability to execute local Python code using the affected library. A fix is available in version 2.7.1-rc1.

Affected products

PyTorch PyTorch < 2.7.1-rc1

Identifiers

CVE-2025-2953
GHSA-3749-ghw9-m3mg
PYSEC-2025-191

Timeline

2025-03-16: disclosed: Issue reported on GitHub
2025-03-30: advisory: GitHub and VulDB advisories published

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.