app.py

import streamlit as st
from transformers import PegasusForConditionalGeneration, PegasusTokenizer, pipeline

# Load the model and tokenizer
model = PegasusForConditionalGeneration.from_pretrained("fatihfauzan26/PEGASUS_liputan6")
tokenizer = PegasusTokenizer.from_pretrained("google/pegasus-cnn_dailymail")

# Initialize the summarization pipeline
summarizer = pipeline("summarization", model=model, tokenizer=tokenizer)

# Streamlit interface
st.title("Summarization App using PEGASUS")

# Input article for summarization
sample_article = st.text_area('Enter the article you want to summarize', height=300)

if sample_article:
    # Generate summary
    input_ids = tokenizer.encode(sample_article, return_tensors='pt')
    summary_ids = model.generate(input_ids,
                                 min_length=30,
                                 max_length=128,
                                 num_beams=8,
                                 repetition_penalty=2.0,
                                 length_penalty=0.8,
                                 early_stopping=True,
                                 no_repeat_ngram_size=2,
                                 use_cache=True,
                                 do_sample=True,
                                 temperature=1.2,
                                 top_k=50,
                                 top_p=0.95)

    # Decode the summary
    summary_text = tokenizer.decode(summary_ids[0], skip_special_tokens=True)

    # Display results
    st.subheader("Summary")
    st.write(summary_text)