EDSD: Entropy-Driven Design for Faster Speculative Decoding.
Longkai Cheng, Ximing Wang, Jiangcai Zhu, Kailai Shao, Chao Chen, Haixiang Hu
Browse the full ACL paper archive.
Longkai Cheng, Ximing Wang, Jiangcai Zhu, Kailai Shao, Chao Chen, Haixiang Hu
Browse the full ACL paper archive.