Xavi99/11chatbot

1import gradio as gr2from huggingface_hub import InferenceClient3 4client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")5 6def respond(message, history, system_message, max_tokens, temperature, top_p):7    messages = [{"role": "system", "content": system_message}]8 9    for val in history:10        if val[0]:11            messages.append({"role": "user", "content": val[0]})12        if val[1]:13            messages.append({"role": "assistant", "content": val[1]})14 15    messages.append({"role": "user", "content": message})16 17    response = ""18 19    for message in client.chat_completion(20        messages,21        max_tokens=max_tokens,22        stream=True,23        temperature=temperature,24        top_p=top_p,25    ):26        token = message.choices[0].delta.content27        response += token28        yield response29 30demo = gr.ChatInterface(31    fn=respond,32    additional_inputs=[33        gr.Textbox(lines=2, label="System Message", placeholder="You are a helpful assistant."),34        gr.Slider(0, 1024, value=256, step=1, label="Max Tokens"),35        gr.Slider(0, 1, value=0.7, step=0.01, label="Temperature"),36        gr.Slider(0, 1, value=0.9, step=0.01, label="Top-p")37    ],38    title="Chat with Zephyr-7b",39    description="Chatbot powered by Hugging Face Inference API."40)41 42demo.launch()  # Remove or set share=False43 44