app.py

import torch
from transformers import pipeline
import gradio as gr

import os
os.environ["HF_HUB_DOWNLOAD_TIMEOUT"] = "300"

asr_pipe = pipeline("automatic-speech-recognition", model="openai/whisper-tiny.en", chunk_length_s=30)
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")

def transcript_and_summarize(audio_file):
    transcript = asr_pipe(audio_file)["text"]
    summary = summarizer(transcript, max_length=100, min_length=30, do_sample=False)[0]["summary_text"]
    return f"**Transcription:**\n{transcript}\n\n**Summary (Notes):**\n{summary}"

iface = gr.Interface(
    fn=transcript_and_summarize, 
    inputs=gr.Audio(sources=["upload"], type="filepath"), 
    outputs=gr.Textbox(), 
    title="Audio Transcription & Note-Making",
    description="Upload an audio file to transcribe and get summarized notes."
)

iface.launch()