CVE-2026-1839

HIGH

Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading in huggingface/transformers

Title source: cna
STIX 2.1

Description

A vulnerability in the HuggingFace Transformers library, specifically in the `Trainer` class, allows for arbitrary code execution. The `_load_rng_state()` method in `src/transformers/trainer.py` at line 3059 calls `torch.load()` without the `weights_only=True` parameter. This issue affects all versions of the library supporting `torch>=2.2` when used with PyTorch versions below 2.6, as the `safe_globals()` context manager provides no protection in these versions. An attacker can exploit this vulnerability by supplying a malicious checkpoint file, such as `rng_state.pth`, which can execute arbitrary code when loaded. The issue is resolved in version v5.0.0rc3.

Scores

CVSS v3 7.8
EPSS 0.0031
EPSS Percentile 21.9%
Attack Vector LOCAL
CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

CISA SSVC

Vulnrichment
Exploitation poc
Automatable no
Technical Impact total

Details

CWE
CWE-502
Status published
Products (4)
huggingface/huggingface/transformers unspecified - v5.0.0rc3
huggingface/transformers 5.0.0 rc0 (3 CPE variants)
huggingface/transformers < 5.0.0
pypi/transformers 0 - 5.0.0rc3PyPI
Published Apr 07, 2026
Tracked Since Apr 07, 2026