vLLM denial of service in pre-tokenized prompt validation
vllm-project vllm
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| vllm PyPI |
|
|
Original description
vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py computes max_token_id with an off-by-one upper bound, allowing a pre-tokenized prompt ID equal to the vocabulary size to pass validation. For a tensor-parallel size of one and an exact multiple-of-64 vocabulary, VocabParallelEmbedding.forward does not mask that value before F.embedding uses it, causing a CUDA device-side assertion and terminating the engine. This issue is fixed in version 0.32.0.
From the CVE record · GitHub_M