README.md

---
title: Accent Classifier
emoji: 🎙️
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.32.0
app_file: app.py
pinned: false
---

# Accent Classifier 🎙️

This Gradio app downloads a public YouTube or Vimeo video, extracts its audio, and classifies the speaker (as a proxy for accent) using a Hugging Face model.

1. Paste a public YouTube or Vimeo URL.  
2. The app uses `yt-dlp` to grab the audio.  
3. `ffmpeg` converts it to a 16 kHz WAV file.  
4. A Hugging Face pipeline (`superb/wav2vec2-base-superb-sid`) predicts a speaker ID and confidence.  
5. Temporary files are cleaned up automatically.

> **Note:** You must have `ffmpeg` installed and on your PATH. If `yt-dlp` is missing, the script will install it at runtime.