speech-tokens

Tag

Cards List
#speech-tokens

CoSTA: Compressed Speech Token Adaptation for Efficient Spoken Dialogue Models

arXiv cs.CL ↗ · 2026-10-16 Cached

本文提出 CoSTA,一个兼容性保留的后置框架,通过均值初始化的词表扩展、发音适配与对话适配,将生成词元空间与合成词元空间解耦,在保留原始语音 tokenizer 和解码器的同时实现约 1.7 倍的语音词元生成速度与更低延迟,且不损失响应质量。

0 favorites 0 likes
#speech-tokens

Using the Mimi codec for metalinguistic representations

arXiv cs.CL ↗ · 2026-08-18 Cached

This paper investigates the Mimi codec's semantic tokens, showing they map to phonetic units like quadphones and triphones, and discusses implications for speech tokenization in AI models.

0 favorites 0 likes
← Back to home

Submit Feedback