app.py

import streamlit as st
from transformers import pipeline

# Set up the title and description of the app
st.title("My LLM Model: Dementia Knowledge Assistant")
st.markdown("""
This app uses a fine-tuned **Large Language Model (LLM)** to answer questions about dementia. 
Simply input your query, and the model will provide contextually relevant answers!
""")

# Load the model
@st.cache_resource
def load_qa_pipeline():
    model_name = "rohitashva/dementia--chatbot-llm-model"  # Replace with your Hugging Face model repo name
    qa_pipeline = pipeline("text-generation", model=model_name)
    return qa_pipeline

qa_pipeline = load_qa_pipeline()

# Input field for user query
st.header("Ask a Question")
question = st.text_input("Enter your question about dementia (e.g., 'What are the symptoms of early-stage dementia?'):")

# Context input for retrieval
st.text_area(
    "Provide additional context (optional):",
    placeholder="Include any relevant context about dementia here to get better results.",
    key="context"
)

if st.button("Get Answer"):
    if not question:
        st.error("Please enter a question!")
    else:
        # Call the QA pipeline
        with st.spinner("Generating response..."):
            result = qa_pipeline({
                "question": question,
                "context": st.session_state.context
            })
            answer = result.get("answer", "I don't know.")
            confidence = result.get("score", 0.0)

        # Display the result
        st.subheader("Answer")
        st.write(answer)

        st.subheader("Confidence Score")
        st.write(f"{confidence:.2f}")

# Footer
st.markdown("---")
st.markdown("Deployed on **Hugging Face Spaces** using [Streamlit](https://streamlit.io/).")