🦔 MOSS Audio Tokenizer — Nano

OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano · CNN-free causal transformer · 12.5 Hz frame rate · 32-layer RVQ · 3M hrs training data

📂 Upload Audio

RVQ layers used for decode (fewer = lower quality / lower bitrate)

1 32

🔊 Reconstructed Audio

Per-RVQ-layer token indices