无法使用assign\u add更新tensorflow变量

我正在尝试用SSU隐藏单元制作RBM，我必须更新标准偏差

我这样定义变量：

    def _build_model(self):
        with tf.device('/gpu:0'):
            with self.graph.as_default():
...
                with tf.variable_scope("visible_layer"):
                    self.v_clamp = tf.placeholder(name = "v_in", dtype = tf.float32, shape=[self.batch_size, self.n_visibles])
                    self.bv = tf.get_variable(name = "b_v", dtype = tf.float32, shape=[self.n_visibles], initializer=tf.random_uniform_initializer(maxval=0.01,minval=-0.01))

                self.stddev = tf.get_variable(name = "stddev", dtype = tf.float32, shape = [1], initializer = tf.constant_initializer(float(self.stddev_)))

...

                with tf.variable_scope("update_weights"):
                    self.optimizer = self.update_weights()

....

    def update_weights(self):
        with self.graph.as_default():
            with tf.device('/gpu:0'):
            ...
            with tf.variable_scope("calc_deltas"):
            ...
                ##UPDATE STDDEV
                delta_stddev = tf.multiply((2)/(self.stddev**3),
                                           tf.subtract(tf.reduce_sum(tf.pow(tf.subtract(self.v_clamp,self.bv),2)),
                                                       tf.reduce_sum(tf.pow(tf.subtract(v_free,self.bv),2))))
            #self.stddev.assing_add(delta_stddev)
            self.stddev.assign_add(tf.constant(0.1,shape=[1]))

            return self.stddev

    def train_model(self):
        with tf.Session(graph=self.graph) as session:
            session.run(tf.global_variables_initializer())#Now all variables should be initialized.
            print("Uninitialized variables: ", session.run(tf.report_uninitialized_variables())) #Just to check, should print nothing

            print("Training for ", self.n_steps)
            for step in range(self.n_steps):

                feed_train = self._create_feed_dict(self.X_train,step)
                feed_test = self._create_feed_dict(self.X_test,step)

                print(session.run(self.optimizer, feed_dict = {self.v_clamp: feed_train}))

其中stddev_u具有初始值

我的更新功能如下：

    def _build_model(self):
        with tf.device('/gpu:0'):
            with self.graph.as_default():
...
                with tf.variable_scope("visible_layer"):
                    self.v_clamp = tf.placeholder(name = "v_in", dtype = tf.float32, shape=[self.batch_size, self.n_visibles])
                    self.bv = tf.get_variable(name = "b_v", dtype = tf.float32, shape=[self.n_visibles], initializer=tf.random_uniform_initializer(maxval=0.01,minval=-0.01))

                self.stddev = tf.get_variable(name = "stddev", dtype = tf.float32, shape = [1], initializer = tf.constant_initializer(float(self.stddev_)))

...

                with tf.variable_scope("update_weights"):
                    self.optimizer = self.update_weights()

....

    def update_weights(self):
        with self.graph.as_default():
            with tf.device('/gpu:0'):
            ...
            with tf.variable_scope("calc_deltas"):
            ...
                ##UPDATE STDDEV
                delta_stddev = tf.multiply((2)/(self.stddev**3),
                                           tf.subtract(tf.reduce_sum(tf.pow(tf.subtract(self.v_clamp,self.bv),2)),
                                                       tf.reduce_sum(tf.pow(tf.subtract(v_free,self.bv),2))))
            #self.stddev.assing_add(delta_stddev)
            self.stddev.assign_add(tf.constant(0.1,shape=[1]))

            return self.stddev

    def train_model(self):
        with tf.Session(graph=self.graph) as session:
            session.run(tf.global_variables_initializer())#Now all variables should be initialized.
            print("Uninitialized variables: ", session.run(tf.report_uninitialized_variables())) #Just to check, should print nothing

            print("Training for ", self.n_steps)
            for step in range(self.n_steps):

                feed_train = self._create_feed_dict(self.X_train,step)
                feed_test = self._create_feed_dict(self.X_test,step)

                print(session.run(self.optimizer, feed_dict = {self.v_clamp: feed_train}))

注释的行是我尝试过的东西

我是这样训练的：

    def _build_model(self):
        with tf.device('/gpu:0'):
            with self.graph.as_default():
...
                with tf.variable_scope("visible_layer"):
                    self.v_clamp = tf.placeholder(name = "v_in", dtype = tf.float32, shape=[self.batch_size, self.n_visibles])
                    self.bv = tf.get_variable(name = "b_v", dtype = tf.float32, shape=[self.n_visibles], initializer=tf.random_uniform_initializer(maxval=0.01,minval=-0.01))

                self.stddev = tf.get_variable(name = "stddev", dtype = tf.float32, shape = [1], initializer = tf.constant_initializer(float(self.stddev_)))

...

                with tf.variable_scope("update_weights"):
                    self.optimizer = self.update_weights()

....

    def update_weights(self):
        with self.graph.as_default():
            with tf.device('/gpu:0'):
            ...
            with tf.variable_scope("calc_deltas"):
            ...
                ##UPDATE STDDEV
                delta_stddev = tf.multiply((2)/(self.stddev**3),
                                           tf.subtract(tf.reduce_sum(tf.pow(tf.subtract(self.v_clamp,self.bv),2)),
                                                       tf.reduce_sum(tf.pow(tf.subtract(v_free,self.bv),2))))
            #self.stddev.assing_add(delta_stddev)
            self.stddev.assign_add(tf.constant(0.1,shape=[1]))

            return self.stddev

    def train_model(self):
        with tf.Session(graph=self.graph) as session:
            session.run(tf.global_variables_initializer())#Now all variables should be initialized.
            print("Uninitialized variables: ", session.run(tf.report_uninitialized_variables())) #Just to check, should print nothing

            print("Training for ", self.n_steps)
            for step in range(self.n_steps):

                feed_train = self._create_feed_dict(self.X_train,step)
                feed_test = self._create_feed_dict(self.X_test,step)

                print(session.run(self.optimizer, feed_dict = {self.v_clamp: feed_train}))

问题是，其他变量，即向量（如self.bv）被正确更新，但这个变量（stddev）始终等于初始值

---

我不知道我做错了什么，这是因为调用

tf.assign\u add

方法时，您没有运行在TensorFlow图中定义的

assign\u add

操作

import tensorflow as tf
v = tf.get_variable('t', shape=[], initializer=tf.constant_initializer(0.))

op = tf.assign_add(v, 1)

with tf.Session() as session:
    session.run(tf.global_variables_initializer())
    print(session.run(v)) # print 0. 
    print(session.run(op)) # print 1. as you just ran the `assign_add` operation
    print(session.run(v)) # print 1. as `v` has been incremented.

编辑：

在您的情况下，您可以做的是：

def update_weights(self):
    ...
    return self.stddev.assing_add(delta_stddev)

---

这样，您的方法将返回

op

，它实际上更新了您的

self.stdv

变量。

谢谢您的回答！但我不明白。调用优化器时我没有运行它吗？实际上没有：当运行

self.stddev.assign\u add（tf.constant（0.1，shape=[1]））时，您只在TensorFlow图中定义了op，但从未显式运行过它。您可以做的是让
update\u weights
返回
self.stddev.assing\u add（delta\u stddev）
。