Update app.py

app.py

@@ -25,12 +25,13 @@ def generate(
     message: str,
     chat_history: list[tuple[str, str]],
     max_new_tokens: int = 1024,
-    temperature: float = 0.06,
-    top_p: float = 0.95,
-    top_k: int = 40,
-    repetition_penalty: float = 1.2,
+    temperature: float = 1,
+    top_p: float = 1.,
+    top_k: int = 50,
+    do_sample: bool = False,
+    repetition_penalty: float = 1.,
 ) -> Iterator[str]:
-    conversation = []
+    conversation = [{"role": "system", "content": ""}]
     for user, assistant in chat_history:
         conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
     conversation.append({"role": "user", "content": message})
@@ -46,7 +47,7 @@ def generate(
         {"input_ids": input_ids},
         streamer=streamer,
         max_new_tokens=max_new_tokens,
-        do_sample=True,
+        do_sample=do_sample,
         top_p=top_p,
         top_k=top_k,
         temperature=temperature,
@@ -102,8 +103,12 @@ chat_interface = gr.ChatInterface(
             minimum=1.0,
             maximum=2.0,
             step=0.05,
-            value=1.2,
+            value=1.,
         ),
+        gr.Checkbox(
+            label="Do sample",
+            value=False,
+        )
     ],
     examples=[
         ["""Vraagje: welk woord hoort er niet in dit rijtje thuis: "auto, vliegtuig, geit, bus"?"""],