Suggestions search

All statuses

Filter by:

With package: python313Packages.tf-keras

Found 2 matching suggestions

View:

Compact

Detailed

Untriaged

Permalink CVE-2026-1669

created 2 months, 1 week ago

Arbitrary File Read in Keras via HDF5 External Datasets

Arbitrary file read in the model loading mechanism (HDF5 integration) in Keras versions 3.0.0 through 3.13.1 on all supported platforms allows a remote attacker to read local files and disclose sensitive information via a crafted .keras model file utilizing HDF5 external dataset references.

References

https://github.com/google/security-research/security/advisories

Affected products

Keras

<3.13.1

Matching in nixpkgs

pkgs.python312Packages.keras

Multi-backend implementation of the Keras API, with support for TensorFlow, JAX, and PyTorch

nixos-unstable 3.11.1
nixos-25.11 3.12.0
- nixpkgs-25.11-darwin 3.12.0

pkgs.python313Packages.keras

Multi-backend implementation of the Keras API, with support for TensorFlow, JAX, and PyTorch

nixos-unstable -
- nixpkgs-unstable 3.13.1
- nixos-unstable-small 3.13.1
nixos-25.11 3.12.0
- nixpkgs-25.11-darwin 3.12.0

pkgs.python312Packages.tf-keras

Deep learning for humans

nixos-unstable 2.19.0
nixos-25.11 2.20.1
- nixpkgs-25.11-darwin 2.20.1

pkgs.python313Packages.tf-keras

Deep learning for humans

nixos-unstable -
- nixpkgs-unstable 2.20.1
- nixos-unstable-small 2.20.1
nixos-25.11 2.20.1
- nixpkgs-25.11-darwin 2.20.1

pkgs.pkgsRocm.python3Packages.keras

Multi-backend implementation of the Keras API, with support for TensorFlow, JAX, and PyTorch

nixos-unstable -
- nixpkgs-unstable 3.13.1
- nixos-unstable-small 3.13.1
nixos-25.11 3.12.0
- nixpkgs-25.11-darwin 3.12.0

Package maintainers

@GaetanLepage Gaetan Lepage <gaetan@glepage.com>

Untriaged

Permalink CVE-2026-0897

created 3 months ago

Denial of Service in Keras via Excessive Memory Allocation in HDF5 Metadata

Allocation of Resources Without Limits or Throttling in the HDF5 weight loading component in Google Keras 3.0.0 through 3.13.0 on all platforms allows a remote attacker to cause a Denial of Service (DoS) through memory exhaustion and a crash of the Python interpreter via a crafted .keras archive containing a valid model.weights.h5 file whose dataset declares an extremely large shape.