From 623e2ed29f70dec5af892871f0307655e348c8a9 Mon Sep 17 00:00:00 2001 From: Nick Hill Date: Mon, 24 Mar 2025 15:58:59 -0700 Subject: [PATCH] [BugFix][V1] Quick fix for min_tokens with multiple EOS (#15407) Signed-off-by: Nick Hill --- vllm/sampling_params.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/vllm/sampling_params.py b/vllm/sampling_params.py index 9b474a37b9..584320e76c 100644 --- a/vllm/sampling_params.py +++ b/vllm/sampling_params.py @@ -369,8 +369,9 @@ class SamplingParams( self.top_k = -1 self.min_p = 0.0 self._verify_greedy_sampling() + # eos_token_id is added to this by the engine - self._all_stop_token_ids = set(self.stop_token_ids) + self._all_stop_token_ids.update(self.stop_token_ids) def _verify_args(self) -> None: if not isinstance(self.n, int):