TF does not support an arbitrary padding in conv2d function, so if you need some padding that is not supported, use tf.pad(). Luckily for our input the ...
確定! 回上一頁