Dynamic Quantization

Dynamic Quantization is a quantization method that dynamically determines the scaling and shifting parameters for weight and activation tensors during inference, distinguishing itself from static quantization approaches by adapting these parameters to the input data distribution on a per-layer or per-tensor basis.

161

Publications

6.9K

Citations

407

Authors

196

Institutions

Publications per year

2017–2026

75

Authors

407

  1. XC
    Xiao‐Heng Chang

    Bohai University, China

    9 publications · 498 citations

    9

    h-index

  2. TS
    Toshiharu Sugie

    Kyoto University, Japan

    8 publications · 440 citations

    8

    h-index

  3. SA
    Shun‐ichi Azuma

    Kyoto University, Japan

    8 publications · 440 citations

    8

    h-index

  4. ZW
    Zidong Wang

    Brunel University of London, United Kingdom

    6 publications · 394 citations

    6

    h-index

  5. NM
    Nasser M. Nasrabadi

    DEVCOM Army Research Laboratory, United States

    4 publications · 104 citations

    4

    h-index

1–5 of 407

Venues

  1. IEEE Transactions on Automatic Control

    16 publications · 1.2K citations

    13

    h-index

  2. International Journal of Robust and Nonlinear Control

    IJRCEA · 5 publications · 256 citations

    5

    h-index

  3. Automatica

    4 publications · 361 citations

    4

    h-index

  4. IEEE Transactions on Signal Processing

    4 publications · 128 citations

    4

    h-index

  5. IEEE Transactions on Image Processing

    4 publications · 134 citations

    4

    h-index

1–5