Allocation of Resources Without Limits or Throttling Affecting keras package, versions [,3.15.0)
Threat Intelligence
Snyk has a proof-of-concept or detailed explanation of how to exploit this vulnerability.
The probability is the direct output of the EPSS model, and conveys an overall sense of the threat of exploitation in the wild. The percentile measures the EPSS probability relative to all known EPSS scores. Note: This data is updated daily, relying on the latest available EPSS model version. Check out the EPSS documentation for more details.
Do your applications use this vulnerable package?
In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.
Test your applicationsSnyk Learn
Learn about Allocation of Resources Without Limits or Throttling vulnerabilities in an interactive lesson.
Start learning- Snyk IDSNYK-PYTHON-KERAS-18601821
- published10 Aug 2026
- disclosed10 Aug 2026
- creditthesecguy45
Introduced: 10 Aug 2026
NewCVE-2026-12570 (opens in a new tab)How to fix?
Upgrade keras to version 3.15.0 or higher.
Overview
keras is a Keras is a high-level neural networks API for Python..
Affected versions of this package are vulnerable to Allocation of Resources Without Limits or Throttling via the safe_get_h5_dataset function in keras/src/saving/saving_lib.py. An attacker can trigger an out-of-memory crash by supplying a malicious .keras or .weights.h5 file to keras.models.load_model() or load_weights() that declares a huge HDF5 dataset shape while storing almost no data on disk. The loader materializes the dataset according to its declared shape without checking the on-disk storage size, so a crafted chunked, compressed, fill-value-only dataset can force unbounded memory allocation and terminate the process. This breaks machine-learning jobs that load untrusted models or weights, causing service disruption and failed model ingestion.