diff --git a/configs/kb_config.py.example b/configs/kb_config.py.example
index c9ff431..6be2848 100644
--- a/configs/kb_config.py.example
+++ b/configs/kb_config.py.example
@@ -21,8 +21,10 @@ OVERLAP_SIZE = 50
 # 知识库匹配向量数量
 VECTOR_SEARCH_TOP_K = 3
 
-# 知识库匹配相关度阈值，取值范围在0-1之间，SCORE越小，相关度越高，取到1相当于不筛选，建议设置在0.5左右
-SCORE_THRESHOLD = 1
+# 知识库匹配的距离阈值，取值范围在0-1之间，SCORE越小，距离越小从而相关度越高，
+# 取到1相当于不筛选，实测bge-large的距离得分大部分在0.01-0.7之间，
+# 相似文本的得分最高在0.55左右，因此建议针对bge设置得分为0.6
+SCORE_THRESHOLD = 0.6
 
 # 默认搜索引擎。可选：bing, duckduckgo, metaphor
 DEFAULT_SEARCH_ENGINE = "duckduckgo"
diff --git a/startup.py b/startup.py
index 3bb508a..11f60b6 100644
--- a/startup.py
+++ b/startup.py
@@ -128,7 +128,7 @@ def create_model_worker_app(log_level: str = "INFO", **kwargs) -> FastAPI:
             args.conv_template = None
             args.limit_worker_concurrency = 5
             args.no_register = False
-            args.num_gpus = 4 # vllm worker的切分是tensor并行，这里填写显卡的数量
+            args.num_gpus = 1 # vllm worker的切分是tensor并行，这里填写显卡的数量
             args.engine_use_ray = False
             args.disable_log_requests = False