📄 Scaling Transformers for End-to-End Discrete Audio Tokenization ✅ 6.0/10 | 前50% | arxiv ← 返回 2026-05-23 语音/音乐/音频论文速递