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.

Log in or Sign Up to review the conditions and access this model content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

NeMo vLLM tokenizer Hydra instantiate RCE PoC

This PoC demonstrates that nemo.export.vllm.model_config.NemoModelConfig instantiates the model-controlled tokenizer._target_ from context/model.yaml before validating that the tokenizer target is a Hugging Face tokenizer.

The malicious NeMo2-style model artifact is:

  • malicious_nemo2_ckpt/context/model.yaml

The tokenizer._target_ uses Hydra to call pydoc.pipepager, which executes a local command and creates nemo_vllm_marker.txt during model config loading.

Run:

python build_and_verify.py

Expected result:

  • marker_created=true
  • marker_text="nemo_vllm_modelconfig_rce"
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support