app.py

import streamlit as st
import transformers
import numpy as np

# Load the pre-trained model
model1 = transformers.pipeline("text2text-generation", model="bigscience/T0pp")
model2 = transformers.pipeline("text2text-generation", model="google/flan-t5-xxl")
model3 = transformers.pipeline("text2text-generation", model="google/flan-t5-xl")
model4 = transformers.pipeline("text2text-generation", model="tuner007/pegasus_paraphrase")
model5 = transformers.pipeline("text2text-generation", model="tuner007/pegasus_paraphrase")

# Define the Streamlit app
def main():
    st.title("Topic Modeling with Hugging Face")
    text = st.text_area("Enter some text to generate topics", height=200)

    if st.button("Generate Topics"):
        # Generate topics
        topics1 = model1(text, max_length=50, do_sample=True, num_beams=5, temperature=0.7)
        topics2 = model2(text, max_length=50, do_sample=True, num_beams=5, temperature=0.7)
        topics3 = model3(text, max_length=50, do_sample=True, num_beams=5, temperature=0.7)
        topics4 = model4(text, max_length=50, do_sample=True, num_beams=5, temperature=0.7)
        topics5 = model5(text, max_length=50, do_sample=True, num_beams=5, temperature=0.7)

        # Print topics
        st.write("Top 5 topics:")
        for i in range(5):
            st.write(f"{i+1}. {topics1[i]['generated_text']}")
            st.write(f"{i+1}. {topics2[i]['generated_text']}")
            st.write(f"{i+1}. {topics3[i]['generated_text']}")
            st.write(f"{i+1}. {topics4[i]['generated_text']}")
            st.write(f"{i+1}. {topics5[i]['generated_text']}")

if __name__ == "__main__":
    main()