GHSA-59p9-h35m-wg4g

Hugging Face Transformers is vulnerable to ReDoS through its MarianTokenizer

At a glance

Published
Updated
CVE
CVE-2025-6638
Aliases
PYSEC-2026-1981
CVSS_V3
CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L

Description

A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer's `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service.

Affected packages

PackageVersions
PyPI · transformersintroduced 0, fixed 4.53.0

References

Sources

SourceLicence
OSV.devCC-BY 4.0 — GitHub Advisory Database