如何使用tf.placeholder输入tf.unstack的参数？

Question

tf.unstack具有参数num，该参数必须为整数才能确定输出将具有多少行。

    tf.unstack(
     value,
     num=None,
     axis=0,
     name='unstack'
    )

我需要使用tf.placeholder输入以下值：

self.num = tf.placeholder(tf.int32, shape = ())

但是由于self.num是张量，所以出现此错误。有什么建议吗？

TypeError: Expected int for argument 'num' not <tf.Tensor 'inputs/Placeholder_2:0' shape=() dtype=int32>.

这是与代码相关的部分：

class PGNetwork:
def __init__(self, state_size, action_size, learning_rate):
    self.state_size = state_size
    self.action_size = action_size
    self.learning_rate = learning_rate
    self.weights = [
                tf.get_variable('wd1', shape = weight_shapes[0], dtype=tf.float64, initializer=tf.contrib.layers.xavier_initializer(dtype=tf.float64)), 
                tf.get_variable('wd2', shape = weight_shapes[1], dtype=tf.float64, initializer=tf.contrib.layers.xavier_initializer(dtype=tf.float64)), 
                tf.get_variable('wd3', shape = weight_shapes[2], dtype=tf.float64, initializer=tf.contrib.layers.xavier_initializer(dtype=tf.float64)), 
                tf.get_variable('bd1', shape = weight_shapes[3], dtype=tf.float64, initializer=tf.contrib.layers.xavier_initializer(dtype=tf.float64)),
                tf.get_variable('bd2', shape = weight_shapes[4], dtype=tf.float64, initializer=tf.contrib.layers.xavier_initializer(dtype=tf.float64)),
                tf.get_variable('bd3', shape = weight_shapes[5], dtype=tf.float64, initializer=tf.contrib.layers.xavier_initializer(dtype=tf.float64)),
                ]


    with tf.name_scope("inputs"):
        self.inputs_ = tf.placeholder(tf.float64, [None, self.state_size], name="inputs_")
        self.actions = tf.placeholder(tf.int32, [None, self.action_size], name="actions")
        self.discounted_episode_rewards_ = tf.placeholder(tf.float64, [None,], name="discounted_episode_rewards")
        self.flat_multiplier_tensor = tf.placeholder(tf.float64, shape = [None])
        self.parameters = tf.placeholder(tf.float64, shape = [None])
        self. num = tf.placeholer(tf.int32, shape = ())



    with tf.name_scope('PGNetwork'):    
        self.logits = policy(self.inputs_, self.weights)


    with tf.name_scope("softmax"):
        self.action_distribution = tf.nn.softmax(self.logits)
        #self.variables = [v for v in tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES)]
        self.var_list = tf.trainable_variables()
        self.trainable_vars = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES)
        self.var_names = [var.name for var in self.trainable_vars]

    with tf.name_scope("sample_gradient"):

        self.split_inputs = tf.unstack(self.inputs_, num = self.num, axis=0)
        self.split_actions = tf.unstack(self.actions, num = self.num, axis = 0)
        self.intermediate = [tf.expand_dims(self.split_inputs[i], 0) for i in range(self.num)]

当我尝试使用sess.run时，出现错误。