Update app.py

app.py

@@ -164,7 +164,7 @@ def process_history(history: list[dict]) -> list[dict]:
 
 
 @spaces.GPU(duration=120)
-def run(message: dict, history: list[dict], ignored_system_prompt: str = "", max_new_tokens: int = 512) -> Iterator[str]:
+def run(message: dict, history: list[dict]) -> Iterator[str]:
     if not validate_media_constraints(message, history):
         yield ""
         return
@@ -186,7 +186,7 @@ def run(message: dict, history: list[dict], ignored_system_prompt: str = "", max
     generate_kwargs = dict(
         inputs,
         streamer=streamer,
-        max_new_tokens=max_new_tokens,
+        max_new_tokens=2048,
     )
     t = Thread(target=model.generate, kwargs=generate_kwargs)
     t.start()
@@ -225,10 +225,6 @@ demo = gr.ChatInterface(
     chatbot=gr.Chatbot(type="messages", scale=1, allow_tags=["image"]),
     textbox=gr.MultimodalTextbox(file_types=["image", ".mp4"], file_count="multiple", autofocus=True),
     multimodal=True,
-    additional_inputs=[
-        gr.Textbox(label="System Prompt", value="You are a helpful assistant."),
-        gr.Slider(label="Max New Tokens", minimum=100, maximum=2000, step=10, value=700),
-    ],
     stop_btn=False,
     title="ChatFinanz",
     examples=examples,