app.py

import torch
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
import streamlit as st




# Model name
model_name = "YasirAbdali/bart-summarization" # Replace with the path to your fine-tuned model or Hugging Face model ID

# Load tokenizer and model
try:
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
    st.write("Model and tokenizer loaded successfully.")
except Exception as e:
    st.error(f"Error loading model or tokenizer: {e}")
    st.stop()

# Streamlit app
st.title("Summary Generator")

# User input
topic = st.text_area("Enter text:")
max_length = st.slider("Maximum length of generated text:", min_value=100, max_value=500, value=200, step=50)

if topic:
    # Tokenize input
    try:
        input_ids = tokenizer.encode(topic, return_tensors="pt")
        st.write("Input text tokenized successfully.")
    except Exception as e:
        st.error(f"Error tokenizing input text: {e}")
        st.stop()

    # Generate summary
    try:
        with torch.no_grad():
            output = model.generate(
                input_ids,
                max_length=max_length,
                num_return_sequences=1,
                no_repeat_ngram_size=2,
                top_k=50,
                top_p=0.95,
                temperature=0.7
            )
        st.write("Summary generated successfully.")
    except Exception as e:
        st.error(f"Error generating summary: {e}")
        st.stop()

    # Decode and display generated summary
    try:
        generated_summary = tokenizer.decode(output[0], skip_special_tokens=True)
        st.subheader("Generated Summary:")
        st.markdown(generated_summary)
    except Exception as e:
        st.error(f"Error decoding generated summary: {e}")

    # Option to download the summary
    st.download_button(
        label="Download Summary",
        data=generated_summary,
        file_name="generated_summary.txt",
        mime="text/plain"
    )