Conv2D ( filters=16, kernel_size= (9,9), strides=(2,2), activation='relu', padding='valid', ... we achieved this using the tf.nn.convolution operation.
確定! 回上一頁