在tensorflow元素中重复/平铺数据

Question

我想在张量流或keras中重复数据，您可以使用void print_player_info(Player * const p) { // only prints the player info so no ownership is taken // p is only use within print_player_info cout << "Name is:" << p->getName() << "\n"; } int main() { Game Game1; Game1.addPlayer(std::move(std::make_unique<Player>(0, "Player1")); Game1.addPlayer(std::move(std::make_unique<Player>(0, "Player2")); print_player_info(Game1.playerList[1]->get()); return 0; } 将tf.tile重复到[a, b, c]。但是，是否有任何简单的方法来获取[a, b, c, a, b, c]？

Answer 1

尝试：

import tensorflow as tf

a = tf.constant(['a','b','c'], dtype=tf.string)
result = tf.reshape(tf.stack([a,a],axis=1), shape=(-1,))

with tf.Session() as sess:
    print(sess.run(result))

# print
[b'a' b'a' b'b' b'b' b'c' b'c']

Answer 2

不幸的是，要实现您想要的功能有点麻烦。在这里，您去了：

X = tf.constant([1, 2, 3])
expanded = tf.expand_dims(X, axis=-1)
#[[1]
# [2]
# [3]]
tiled = tf.tile(expanded,[1,2])
# [[1 1]
# [2 2]
# [3 3]]
reshaped = tf.reshape(tiled,[-1]) 
# [1 1 2 2 3 3]