A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 13:30 UTC
[Misc] Validate `stop_token_ids` contents (#17268)
Signed-off-by: Nick Hill <nhill@redhat.com>
Nick Hill committed
Apr 27, 2025 at 20:54 UTC
8262a3e23b4dc8a29793dff0329ca0743373c85f
1 file changed
+4
vllm/sampling_params.py
+4
@@ -437,6 +437,10 @@ class SamplingParams(
437
and self.truncate_prompt_tokens < 1):
438
raise ValueError(f"truncate_prompt_tokens must be >= 1, "
439
f"got {self.truncate_prompt_tokens}")
440
+ assert isinstance(self.stop_token_ids, list)
441
+ if not all(isinstance(st_id, int) for st_id in self.stop_token_ids):
442
+ raise ValueError(f"stop_token_ids must contain only integers, "
443
+ f"got {self.stop_token_ids}.")
444
assert isinstance(self.stop, list)
445
if any(not stop_str for stop_str in self.stop):
446
raise ValueError("stop cannot contain an empty string.")