OneLA: Scaling Linear-Attention Decoding to Large Beams in Generative Recommendation
This work presents OneLA, a linear-attention decoding framework that exploits the shared prompt and short divergent suffixes of GR workloads, and represents all beam states using a single shared prompt-derived state and compact, append-only records of their divergent transitions.
Xiang-Rui Yang, Cheng Peng, Yun-Feng Zhao et al.
· 1 citation