22.9 A 12nm 18.1tflops/w Sparse Transformer Processor with Entropy-Based Early Exit, Mixed-Precision Predication and Fine-Grained Power Management

Thierry Tambe,Jeff Zhang,Coleman Hooper,Tianyu Jia,Paul N. Whatmough,Joseph Zuckerman,Maico Cassel Dos Santos,Erik Jens Loscalzo,Davide Giri,Kenneth Shepard,Luca Carloni,Alexander Rush,David Brooks,Gu-Yeon Wei

IEEE International Solid-State Circuits Conference（2023）

引用 2|浏览9

关键词

algorithmic optimizations,BERT,entropy-based early exit,feed-forward network,fine-grained power management,inference pass,intelligent conversational interfaces,language models,natural language processing,NLP,TFLOP,transformer layers,transformer models,virtual assistants

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要