Padding is just a fancy name of telling: surround your input matrix with some constant. ... res = tf.squeeze(tf.nn.conv2d(image, kernel, [1, 1, 1, 1], ...
確定! 回上一頁