In 3D convolution, the same 3D kernel is applied to overlapping Summary. ... and MaxPooling2D layer is a 3D tensor of shape (height, width, channels).
確定! 回上一頁