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, 23:30 UTC
[Bugfix]Fix search start_index of stop_checker (#13280)
Xu Song committed
Feb 15, 2025 at 13:39 UTC
067fa2255b6687ccaa79391dc9d1a08c7632f605
1 file changed
+1
-1
vllm/engine/output_processor/stop_checker.py
+1
-1
@@ -113,7 +113,7 @@ class StopChecker:
113
stop_string_len = len(stop_str)
114
# Avoid searching already-searched text.
115
stop_index = output_text.find(stop_str,
116
- -new_char_count - stop_string_len)
116
+ 1 - new_char_count - stop_string_len)
117
if stop_index == -1:
118
continue
119