2019/09/17 by Abhisek Kundu, Kundu, Abhisek, Alex Heinecke +17
Computer Science · #Advanced Neural Network Applications #Algorithms and Data Compression #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Natural Language Processing Techniques #Neural and Evolutionary Computing (cs.NE)
paper · pdf · doi:10.48550/arxiv.1909.07729
openalex publication_date 2019/09/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We propose K-TanH, a novel, highly accurate, hardware efficient approximation of popular activation function TanH for Deep Learning. K-TanH consists of parameterized low-precision integer operations, such as, shift and add/subtract (no floating point operation needed) where parameters are stored in very small look-up tables that can fit in CPU registers. K-TanH can work on various numerical formats, such as, Float32 and BFloat16. High quality approximations to other activation functions, e.g., Sigmoid, Swish and GELU, can be derived from K-TanH. Our AVX512 implementation of K-TanH demonstrates >5× speed up over Intel SVML, and it is consistently superior in efficiency over other approximations that use floating point arithmetic. Finally, we achieve state-of-the-art Bleu score and convergence results for training language translation model GNMT on WMT16 data sets with approximate TanH obtained via K-TanH on BFloat16 inputs.