Refa: ollama keep alive issue. (#8216)

### What problem does this PR solve?

#8122

### Type of change

- [x] Refactoring

rag/llm/cv_model.py

@@ -506,7 +506,8 @@ class OllamaCV(Base):
             response = self.client.chat(
                 model=self.model_name,
                 messages=history,
-                options=options
+                options=options,
+                keep_alive=-1
             )
 
             ans = response["message"]["content"].strip()
@@ -536,7 +537,8 @@ class OllamaCV(Base):
                 model=self.model_name,
                 messages=history,
                 stream=True,
-                options=options
+                options=options,
+                keep_alive=-1
             )
             for resp in response:
                 if resp["done"]: