app.py

import streamlit as st
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Set page configuration - MUST be the first Streamlit command
st.set_page_config(
    page_title="EdStem Chatbot Demo",
    page_icon="🤖",
    layout="wide",
)

# Load model and tokenizer
@st.cache_resource
def load_model():
    model_name = "ML-GT/CS4641-7641-finetuned-phi-3-mini-128k-instruct"  # Replace with the model path or identifier
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float32)
    return tokenizer, model

tokenizer, model = load_model()

# App UI
st.title("🤖 EdStem Chatbot Demo")
st.markdown(
    """
    This is a demo of a fine-tuned LLaMA 3.1 model for EdStem course-related Q&A. 
    Enter your questions below and see how the chatbot responds!
    """
)

# Initialize session state
if "messages" not in st.session_state:
    st.session_state.messages = []

# Chat Interface
def add_message(user_message, bot_message):
    """Append user and bot messages to the session state."""
    st.session_state.messages.append({"user": user_message, "bot": bot_message})

# Input box for user to type their message
user_input = st.text_input("Type your message:", placeholder="Ask a course-related question here...")

if user_input:
    # Generate response for the user's input
    inputs = tokenizer(user_input, return_tensors="pt")
    outputs = model.generate(inputs["input_ids"], max_length=200, num_return_sequences=1)
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    
    # Add user input and bot response to the chat
    add_message(user_input, response)
    
    # Clear input box after submission
    st.experimental_rerun()

# Display chat history
st.write("---")
for chat in st.session_state.messages:
    st.markdown(f"**You:** {chat['user']}")
    st.markdown(f"**Bot:** {chat['bot']}")
    st.write("---")