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

# Load the GPT-2 large model and tokenizer
model_name = "gpt2-large"
model = AutoModelForCausalLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

def generate_blogpost(topic):
    try:
        # Prepare input text
        input_text = f"Write a blog post about {topic}:"
        inputs = tokenizer.encode(input_text, return_tensors="pt")
        st.write(f"Input IDs: {inputs}")

        # Generate output
        with torch.no_grad():
            outputs = model.generate(inputs, max_length=500, num_return_sequences=1)
        st.write(f"Output IDs: {outputs}")

        # Decode the generated text
        generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
        return generated_text

    except Exception as e:
        return f"An error occurred: {str(e)}"

# Streamlit UI
st.title("Blog Post Generator")
st.write("Generate a blog post for a given topic using GPT-2 large.")

# Input for the topic
topic = st.text_input("Enter the topic:")

# Generate button
if st.button("Generate"):
    if topic:
        # Generate and display the blog post
        blog_post = generate_blogpost(topic)
        st.write(blog_post)
    else:
        st.write("Please enter a topic to generate a blog post.")