1 paper · 1 filter
Junrui Xiao, Zhikai Li, Lianwei Yang +1
As emerging hardware begins to support mixed bit-width arithmetic computation, mixed-precision quantization is widely used to reduce the complexity of neural networks. However, Vis…