app.py

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

# Initialize the BART model and tokenizer
MODEL_NAME = "facebook/bart-large-cnn"
model = BartForConditionalGeneration.from_pretrained(MODEL_NAME)
tokenizer = BartTokenizer.from_pretrained(MODEL_NAME)

def convert_and_summarize(audio_path: str) -> str:
    # Convert audio to text
    whisper_model = whisper.load_model("base")
    result = whisper_model.transcribe(audio_path)
    transcribed_text = result["text"]
    
    # Summarize the transcribed text
    inputs = tokenizer([transcribed_text], max_length=1024, truncation=True, return_tensors='pt')
    summary_ids = model.generate(inputs['input_ids'])
    summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
    
    return summary

audio_input = gr.inputs.Audio(type="filepath")

# Interface for Gradio
iface = gr.Interface(
    fn=convert_and_summarize,
    inputs=audio_input,
    outputs="text",
    title="Audio-to-Summarized-Text",
    description="Upload an audio here and get a bullet-point summary of its content.",
    theme="Monochrome",
    live=True,
    capture_session=True,
)

iface.launch()