NVIDIA-NeMo / NVIDIA-NeMo/Guardrails

bug: snowflake embedding model not pre-cached in jailbreak detection Dockerfile

Open
#1,648 2 comments 0 reactions 1 assignee View on GitHub

@Pouyanpi is already working on this.

Since Mar 30, 2026.

bug
Dominant language
Python
Stars
7.2k
Forks
842
Avg merge
3d 1h
Merged PRs (30d)
25

Description

just noticed a pre-existing issue, not introduced by this PR. The container downloads the Snowflake/snowflake-arctic-embed-m-long model at runtime on first startup, meaning the container requires internet access to start and the trust_remote_code prompt fires without explicit optin. The Dockerfile pre-downloads GPT2 at build time but not the Snowflake embedding model.

as a follow-up, we should add a similar RUN python -c "..." to both Dockerfiles to pre-cache it during the build.

Originally posted by @Pouyanpi in https://github.com/NVIDIA-NeMo/Guardrails/pull/1596#pullrequestreview-3815086088

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.