GitHub Advisoryadvisory · PyPI

GHSA-9356-575x-2w9m Hugging Face Transformers Regular Expression Denial of Service (ReDoS) vulnerability

MEDIUM Published 2025-08-06 Modified 2025-08-06 Official page ↗
Aliases
CVE-2025-5197
Ecosystem
PyPI
Packages
transformers
Severity
Moderate
CVSS vector
CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L
Weaknesses
CWE-1333
Reviewed
2025-08-06

Affected packages

• PyPI transformers: < 4.53.0 → fixed in 4.53.0

Description

A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the `convert_tf_weight_name_to_pt_weight_name()` function. This function, responsible for converting TensorFlow weight names to PyTorch format, uses a regex pattern `/[^/]*___([^/]*)/` that can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. The vulnerability affects versions up to 4.51.3 and is fixed in version 4.53.0. This issue can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting model conversion processes between TensorFlow and PyTorch formats.

Related

Alias · CVE (1)

  • CVE-2025-5197 A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the… vulnerability

Weakness · CWE (1)

  • CWE-1333 Inefficient Regular Expression Complexity weakness

References