You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

keras-rwkv-tokenizer-poc

Security research test artifact. Do not download or load this file.

This repository contains a deliberately malicious .keras file constructed for security research. The file is a proof-of-concept for an arbitrary code execution chain in keras-hub RWKVTokenizer that bypasses Keras's safe_mode=True default protection.

The artifact is hosted publicly to observe ProtectAI scanner response as part of a responsible disclosure submission to huntr.com's Model File Vulnerability program. Disclosure target: keras-team/keras-hub.

If you are a security researcher, scanner operator, or HuggingFace abuse team member: this artifact is intentionally constructed and disclosed. Loading it will execute Python code in your environment.

Researcher: Richard Askew (@ricaskew on GitHub, Tecnonix on HuggingFace).

Downloads last month: -

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support