import streamlit as st from huggingface_hub import InferenceClient # Initialize the Inference client with the model client = InferenceClient("HuggingFaceH4/zephyr-7b-beta") # Function to generate a response def respond(message, history, 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 = "" # Make the API call and stream the 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 # Streamlit app layout st.title("Zephyr Chatbot") # Textbox for user input user_message = st.text_input("Your message:") # Text area for displaying chat history history = st.session_state.get("history", []) # System message (initialization) system_message = st.text_area("System message", value="You are a friendly Chatbot.") # Sliders for max tokens, temperature, and top-p max_tokens = st.slider("Max new tokens", min_value=1, max_value=2048, value=512, step=1) temperature = st.slider("Temperature", min_value=0.1, max_value=4.0, value=0.7, step=0.1) top_p = st.slider("Top-p (nucleus sampling)", min_value=0.1, max_value=1.0, value=0.95, step=0.05) # Button to send the message if st.button("Send"): # Get the response from the model response_text = "" for text in respond(user_message, history, system_message, max_tokens, temperature, top_p): response_text = text # Update chat history in session state history.append((user_message, response_text)) st.session_state["history"] = history # Display chat history for user_msg, assistant_msg in history: st.write(f"**You:** {user_msg}") st.write(f"**Bot:** {assistant_msg}")