audio-representation-learning
S-SONDO: Lightweight audio embeddings from self-supervised knowledge distillation