use selection

app_local.py

@@ -17,7 +17,7 @@ llm_model = None
 def respond(
     message,
     history: list[tuple[str, str]],
-    model,
+    selected_model,
     system_message,
     max_tokens,
     temperature,
@@ -29,16 +29,17 @@ def respond(
 
     global llm
     global llm_model
-    
-    if llm is None or llm_model != model:
+	
+    # Check if the selected model has changed
+    if llm is None or llm_model != selected_model:
         llm = Llama(
-            model_path=f"models/{model}",
+            model_path=f"models/{selected_model}",
             flash_attn=True,
             n_gpu_layers=81,
             n_batch=1024,
             n_ctx=8192,
         )
-        llm_model = model
+        llm_model = selected_model
 
     provider = LlamaCppPythonProvider(llm)