Update app.py

app.py

@@ -55,7 +55,7 @@ def generate(message, history, system, max_tokens, temperature, top_p, top_k, pe
         top_k=top_k,
         repetition_penalty=penalty, 
         max_tokens=max_tokens,
-        eos_token_id=[151645,151643],
+        stop_token_ids=[151645,151643],
     )
     # generate outputs
     llm = LLM(model=model)
@@ -89,7 +89,14 @@ with gr.Blocks(css=css) as demo:
                 label="Max tokens",
                 render=False,
             ),
-            gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
+            gr.Slider(
+                minimum=0.1, 
+                maximum=1.0, 
+                value=0.7, 
+                step=0.1, 
+                label="Temperature",
+                render=False,
+            ),
             gr.Slider(
                 minimum=0.1,
                 maximum=1.0,