PT-2025-32158 · Hugging Face · Huggingface/Transformers

Published 2025-08-06 · Updated 2025-10-21 · CVE-2025-5197

Positive Technologies (AI Summary)

CVSS v3.1

5.3

Medium

Base vector

AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L

Name of the Vulnerable Software and Affected Versions Hugging Face Transformers versions up to 4.51.3

Description A Regular Expression Denial of Service (ReDoS) vulnerability exists in the convert tf weight name to pt weight name() function of the Hugging Face Transformers library. This function, which converts TensorFlow weight names to PyTorch format, utilizes a regular expression pattern /[^/]* ([^/]*)/ susceptible to exploitation. Crafted input strings can trigger catastrophic backtracking, leading to excessive CPU consumption. This issue can result in service disruption, resource exhaustion, and potential vulnerabilities in API services that rely on model conversion between TensorFlow and PyTorch formats.

Recommendations Update to Hugging Face Transformers version 4.53.0 or later.

Exploit

Fix

DoS

Weakness Enumeration

CWE-1333

Related Identifiers

CVE-2025-5197

GHSA-9356-575X-2W9M

Affected Products

Huggingface/Transformers

References · 12

• 🔥 https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf · Exploit
• https://github.com/huggingface/transformers/commit/944b56000be5e9b61af8301aa340838770ad8a0b · Patch
• https://osv.dev/vulnerability/CVE-2025-5197 · Vendor Advisory
• https://nvd.nist.gov/vuln/detail/CVE-2025-5197 · Security Note
• https://osv.dev/vulnerability/GHSA-9356-575x-2w9m · Vendor Advisory
• https://github.com/huggingface/transformers · Note
• https://github.com/huggingface/transformers/commit/701caef704e356dc2f9331cc3fd5df0eccb4720a · Note
• https://t.me/cveNotify/139646 · Telegram Post
• https://t.me/cveNotify/131063 · Telegram Post
• https://t.me/CVEtracker/29334 · Telegram Post
• https://t.me/cveNotify/131058 · Telegram Post
• https://twitter.com/CVEnew/status/1953072015390158908 · Twitter Post