将张量转换为向量

Question

我有一个形状的张量：

2 x 2 x 3 x 2

即：{玩家、回合、加注次数、采取的行动}

这是一个两人游戏，两轮后结束，最多可以进行 3 次加注，每次加注都包含有关所采取行动的信息。

我想创建一个代表游戏每一步的数组。所以我有关于每个玩家每次加注、每一轮的行动的信息。

有没有可能做这样的事情：

players = array[0, 1]
rounds = array[2, 3]
raises = array[4, 5, 6]
actions = array[7, 8]

niceAndHandyArray = someFunction(players, rounds, raises, actions)
# Which outputs something like this:
[
    [0, [
            2, [
                   4, [7, 8]
                ],
            ...
        ],
    [1, 
         ... 
    ]
]

# And then access it like:
niceAndHandyAcrray[player_index][round_index][raise_index][action_index] = 1

# And pass it to my neural network later on as:
niceAndHandyArray.flatten()

Answer 1

尝试使用reshape:

#Example:
# tensor 't' is [[[1, 1, 1],
#                 [2, 2, 2]],
#                [[3, 3, 3],
#                 [4, 4, 4]],
#                [[5, 5, 5],
#                 [6, 6, 6]]]
# tensor 't' has shape [3, 2, 3]
# pass '[-1]' to flatten 't'
tf.reshape(t, [-1])
#output: [1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5, 6, 6, 6]