Update app.py

app.py

@@ -85,7 +85,7 @@ def generate(
         {"input_ids": input_ids},
         streamer=streamer,
         max_new_tokens=max_new_tokens,
-        do_sample=True,
+        do_sample=False if temperature == 0.0 else True,
         top_p=top_p,
         top_k=top_k,
         temperature=temperature,
@@ -113,31 +113,31 @@ chat_interface = gr.ChatInterface(
         ),
         gr.Slider(
             label="Temperature",
-            minimum=0.1,
+            minimum=0.0,
             maximum=4.0,
             step=0.1,
-            value=0.6,
+            value=0.0,
         ),
         gr.Slider(
             label="Top-p (nucleus sampling)",
             minimum=0.05,
             maximum=1.0,
             step=0.05,
-            value=0.9,
+            value=0.95,
         ),
         gr.Slider(
             label="Top-k",
             minimum=1,
             maximum=1000,
             step=1,
-            value=50,
+            value=20,
         ),
         gr.Slider(
             label="Repetition penalty",
             minimum=1.0,
             maximum=2.0,
             step=0.05,
-            value=1.2,
+            value=1.0,
         ),
     ],
     stop_btn=None,