Attempting to use uninitialized value beta2_power -------TensorFlow報錯

版本：

Python=3.7

TensorFlow=1.14

 

具體代碼：

    init=[tf.global_variables_initializer(), tf.local_variables_initializer()]

    # train optimization
    trainer = tf.train.AdamOptimizer(learning_rate=0.00001, beta1=0.9, beta2=0.999).minimize(train_loss)

 

 

 

 

 

報錯：   Attempting to use uninitialized value beta2_power

 

 

 

 

查找資料：

https://stackoverflow.com/questions/47765595/tensorflow-attempting-to-use-uninitialized-value-beta1-power

 

 

 

得知，  優化器  trainer = tf.train.AdamOptimizer(learning_rate=0.00001, beta1=0.9, beta2=0.999).minimize(train_loss)

需要 局部變量，即 tf.local_variables  ,  在使用該優化器時需要在定義優化器之前定義對局部變量初始化的操作。

 

 

將原代碼改為：

    # train optimization
    trainer = tf.train.AdamOptimizer(learning_rate=0.00001, beta1=0.9, beta2=0.999).minimize(train_loss)

    init=[tf.global_variables_initializer(), tf.local_variables_initializer()]

 

 

 

 

 

 

代碼可正常運行，報錯問題解決。

 

 

 

-------------------------------------------------------------------------------