Home

Description

LLama Factory enables fine-tuning of large language models. Prior to version 1.0.0, a critical vulnerability exists in the `llamafy_baichuan2.py` script of the LLaMA-Factory project. The script performs insecure deserialization using `torch.load()` on user-supplied `.bin` files from an input directory. An attacker can exploit this behavior by crafting a malicious `.bin` file that executes arbitrary commands during deserialization. This issue has been patched in version 1.0.0.

PUBLISHED Reserved 2025-04-24 | Published 2025-05-01 | Updated 2025-05-02 | Assigner GitHub_M




MEDIUM: 6.1CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:U/C:H/I:L/A:L

Problem types

CWE-502: Deserialization of Untrusted Data

Product status

< 1.0.0
affected

References

github.com/...actory/security/advisories/GHSA-f2f7-gj54-6vpv exploit

github.com/...actory/security/advisories/GHSA-f2f7-gj54-6vpv

github.com/...ommit/2989d39239d2f46e584c1e1180ba46b9768afb2a

cve.org (CVE-2025-46567)

nvd.nist.gov (CVE-2025-46567)

Download JSON