Skip to main content
. 2021 Feb 19;15:601109. doi: 10.3389/fnins.2021.601109

FIGURE 13.

FIGURE 13

The learning accuracy based on fixed-point and dynamic fixed-point representations. (A) The learning accuracy based on fixed-point representation of fractional part with different bitwidth. (B) The learning accuracy based on dynamic fixed-point representation with different bitwidth.