import gradio as gr
from huggingface_hub import InferenceClient
client = InferenceClient("http://127.0.0.1:8080")
def respond(
message,
history: list[tuple[str, str]],
system_message,
max_tokens,
temperature,
top_p,
):
messages = [{"role": "system", "content": system_message}]
for val in history:
if val[0]:
messages.append({"role": "user", "content": val[0]})
if val[1]:
messages.append({"role": "assistant", "content": val[1]})
messages.append({"role": "user", "content": message})
response = ""
for message in client.chat_completion(
messages,
max_tokens=max_tokens,
stream=True,
temperature=temperature,
top_p=top_p,
):
token = message.choices[0].delta.content
response += token
yield response
system_prompt = """You are a well-trained AI assistant, your name is Marco-o1. Created by AI Business of Alibaba International Digital Business Group.
## IMPORTANT!!!!!!
When you answer questions, your thinking should be done in <Thought>, and your results should be output in <Output>.
<Thought> should be in English as much as possible, but there are 2 exceptions, one is the reference to the original text, and the other is that mathematics should use markdown format, and the output in <Output> needs to follow the language of the user input.
"""
demo = gr.ChatInterface(
respond,
additional_inputs=[
gr.Textbox(value=system_prompt, label="System message"),
gr.Slider(minimum=1, maximum=2048, value=1, step=1, label="Max new tokens"),
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=0.95,
step=0.05,
label="Top-p (nucleus sampling)",
),
],
)
if __name__ == "__main__":
demo.launch()