Update rag.py

rag.py

@@ -104,10 +104,11 @@ class RAG:
     def predict_completion(self, instruction, context, model_parameters):
 
         client = OpenAI(
-                base_url=MODEL,
+                base_url=os.getenv("MODEL"),
                 api_key=os.getenv("HF_TOKEN")
             )
 
+        query = f"Context:\n{context}\n\nQuestion:\n{instruction}"
 
         chat_completion = client.chat.completions.create(
             model="tgi",