Update

app.py

@@ -34,22 +34,14 @@ model.eval()
 @spaces.GPU(duration=90)
 def generate(
     message: str,
-    chat_history: list[tuple[str, str]],
+    chat_history: list[dict],
     max_new_tokens: int = 1024,
     temperature: float = 0.6,
     top_p: float = 0.9,
     top_k: int = 50,
     repetition_penalty: float = 1.2,
 ) -> Iterator[str]:
-    conversation = []
-    for user, assistant in chat_history:
-        conversation.extend(
-            [
-                {"role": "user", "content": user},
-                {"role": "assistant", "content": assistant},
-            ]
-        )
-    conversation.append({"role": "user", "content": message})
+    conversation = [*chat_history, {"role": "user", "content": message}]
 
     input_ids = tokenizer.apply_chat_template(conversation, add_generation_prompt=True, return_tensors="pt")
     if input_ids.shape[1] > MAX_INPUT_TOKEN_LENGTH:
@@ -126,6 +118,7 @@ demo = gr.ChatInterface(
         ["Write a 100-word article on 'Benefits of Open-Source in AI research'"],
     ],
     cache_examples=False,
+    type="messages",
     description=DESCRIPTION,
     css_paths="style.css",
     fill_height=True,