Padding is the most popular tool for handling this issue. ... from mxnet.gluon import nn npx.set_np() ... Conv2D(1, kernel_size=(5, 3), padding=(2, 1))
確定! 回上一頁