Add modal parameter that is compatible with vllm

app.py

@@ -17,8 +17,8 @@ CHAT_CLIENTS = {
     ),
 }
 
+MODEL_ID = {"Salamandra": "/models/alinia/toxic-sal"}
 MODEL_ARGS = {
-    "model": "default",
     "max_tokens": 256,
     "temperature": 0.3,
     "frequency_penalty": 0.1,
@@ -101,7 +101,9 @@ async def assistant(chat_history, system_prompt, model_name):
 
     try:
         stream = await client.chat.completions.create(
-            **MODEL_ARGS, messages=chat_history
+            **MODEL_ARGS,
+            model=MODEL_ID.get(model_name, "default"),
+            messages=chat_history,
         )
 
         async for chunk in stream:
@@ -152,7 +154,7 @@ with gr.Blocks(title="🦎 Salamandra & Oranguten") as demo:
             with gr.Row():
                 new_chat = gr.Button("New chat")
 
-            response_safety = gr.Label(show_label=False)
+            response_safety = gr.Label(show_label=False, show_heading=False)
 
     # Event Listeners:
     message.submit(user, inputs=[message, chatbot], outputs=[message, chatbot]).then(