README.md

metadata

title: Streaming Vocos Demo
emoji: 🎙️
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.44.1
python_version: '3.10'
app_file: app.py
pinned: false

Streaming Vocos Demo (Gradio)

Upload or record an audio clip, then click Reconstruct to run it through the Streaming Vocos vocoder.

UI layout

• Left: input waveform + input mel-spectrogram
• Right: reconstructed waveform + reconstructed mel-spectrogram

The Space loads the Lightning checkpoint from the Hugging Face model repo and runs:

• mel = model.feature_extractor(audio)
• audio_out = model(mel) (optionally in streaming chunk mode)

Model: warisqr007/StreamingVocos