DOA: Training-Free Decoder-Only Attention Policy for Long-Form Simultaneous Translation with SpeechLLMs 文章

ArXiv CS.CL2026-06-01NEWSen作者: Sara Papi, Luisa Bentivogli
查看原文 →
关系图谱

DOA: Training-Free Decoder-Only Attention Policy for Long-Form Simultaneous Translation with SpeechLLMs · 相关技术

相关技术

Decoder-Only AttentionSpeechLLMslarge language modelsDOASimulST