app.py

import gradio as gr
from transformers import BartTokenizer, BartForConditionalGeneration, BartConfig

# Load tokenizer and model
tokenizer = BartTokenizer.from_pretrained("facebook/bart-large-cnn")
config = BartConfig.from_pretrained("./models/bart-summarizer/checkpoint-8000/config.json")
model_path = "./models/bart-summarizer/checkpoint-8000/"
model = BartForConditionalGeneration.from_pretrained(pretrained_model_name_or_path=model_path, config=config)

# Define summarize function
def summarize(text):
    inputs = tokenizer([text], max_length=1024, return_tensors='pt', truncation=False)
    summary_ids = model.generate(inputs['input_ids'], num_beams=4, min_length=30, max_length=128, early_stopping=True)
    summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)
    return summary

# Create Gradio interface
inputs = gr.Textbox(lines=10, label="Input Text")
outputs = gr.Textbox(label="Summary")

gr.Interface(summarize, inputs, outputs, title="Mail Subject Extraction", description="Get Subject from Email Content").launch()