Enable GPU laer and CORS setting

app/main.py

@@ -20,7 +20,7 @@ app = FastAPI()
 # Set all origins to wildcard for simplicity, but we should limit this in production
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["*"],
+    allow_origins=["https://educonnect-debd1.web.app"],
     allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],

app/utils/chat_rag.py

@@ -92,7 +92,7 @@ class LlamaModelSingleton:
             cls._instance.llm = LlamaCpp(
                 #streaming = True,
                 model_path=model_path,
-                #n_gpu_layers=-1,
+                n_gpu_layers=-1,
                 n_batch=512,
                 temperature=0.1,
                 top_p=1,