MEDIUM 4.3

PYSEC-2026-1984

Transformers Regular Expression Denial of Service (ReDoS) vulnerability

Details

A Regular Expression Denial of Service (ReDoS) vulnerability was identified in the huggingface/transformers library, specifically in the file `tokenization_gpt_neox_japanese.py` of the GPT-NeoX-Japanese model. The vulnerability occurs in the SubWordJapaneseTokenizer class, where regular expressions process specially crafted inputs. The issue stems from a regex exhibiting exponential complexity under certain conditions, leading to excessive backtracking. This can result in high CPU usage and potential application downtime, effectively creating a Denial of Service (DoS) scenario. The affected version is v4.48.1 (latest).

Are you affected?

Enter the version of the package you're using.

Affected packages

PyPI / transformers

Introduced in: 0 Fixed in: 4.50.0

Fix pip install --upgrade 'transformers>=4.50.0'

References

https://nvd.nist.gov/vuln/detail/CVE-2025-1194 [ADVISORY]
https://github.com/huggingface/transformers/commit/92c5ca9dd70de3ade2af2eb835c96215cc50e815 [WEB]
https://github.com/huggingface/transformers [PACKAGE]
https://huntr.com/bounties/86f58dcd-683f-4adc-a735-849f51e9abb2 [WEB]
https://pypi.org/project/transformers [PACKAGE]
https://github.com/advisories/GHSA-fpwr-67px-3qhx [ADVISORY]