THREATOPS
THREAT OPS › Threat News › [GHSA] GHSA-8397-8qf6-f2gm (medium) — vLLM: Phi-4 multimodal audio encoder materializes and retains full attention matrices: a single max-length audio OOM-crashes the vLLM engine

[GHSA] GHSA-8397-8qf6-f2gm (medium) — vLLM: Phi-4 multimodal audio encoder materializes and retains full attention matrices: a single max-length audio OOM-crashes the vLLM engine

medgithub_advisoriesPublished 2026-10-28

GHSA-8397-8qf6-f2gm Severity: medium CVE: CVE-2026-115913

vLLM: Phi-4 multimodal audio encoder materializes and retains full attention matrices: a single max-length audio OOM-crashes the vLLM engine

### Summary

The Phi-4 multimodal audio encoder (`phi4mm_audio.py` / `phi4mm_utils.py`) runs its conformer attention in **eager mode by default**: the official HF config of `microsoft/Phi-4-multimoda

Indicators of compromise

Original source: https://github.com/advisories/GHSA-8397-8qf6-f2gm