Update services/llama_generator.py

services/llama_generator.py

@@ -6,7 +6,9 @@ from datetime import datetime
 import logging
 from config.config import settings
 
-import services.prompt_builder
+from services.prompt_builder import LlamaPromptTemplate
+from services.model_manager import ModelManager
+
 from services.base_generator import BaseGenerator
 
 import asyncio
@@ -46,6 +48,7 @@ class LlamaGenerator(BaseGenerator):
         print(llama_model_name)
         print(prm_model_path)
 
+        self.model_manager = ModelManager()
             
         self.tokenizer = model_manager.load_tokenizer(llama_model_name)  # Add this line to initialize the tokenizer