This is a Streamlit web app that detects the spoken accent from a video file using a pretrained model from Hugging Face (Jzuluaga/accent-id-commonaccent_ecapa). It downloads the video, extracts the audio, and performs accent classification using SpeechBrain.
- 🔗 Accepts a direct video URL (
.mp4) - 🎧 Extracts audio from video
- 🧠 Classifies speaker accent using ECAPA-TDNN model
- 📊 Returns confidence score and prediction
- 🖥️ Interactive UI with Streamlit