tensorflow2.0——可訓練變量

        

 

 

 

 

 

import tensorflow as tf
import numpy as np

###############     tf.Variable(initial value,dtype)    ###############

print('############數字為參數###########')
a = tf.Variable(3)
print('數字為參數a:',a)
print('############列表為參數###########')
a = tf.Variable([1,6])
print('列表為參數a:',a)
print('############np數組為參數###########')
a = tf.Variable(np.array([3,6.0]))
print('np數組為參數a:',a)
print('############張量為參數###########')
a = tf.Variable(tf.constant([[1,1],[2,2],[2,3]]))
print('張量為參數a:',a)
print('a.trainable:',a.trainable)           #   該變量是否可以被訓練
print('type(a):',type(a))
print()
###############     對象名.assign()    ###############
a = tf.Variable([1,2,3])
print('原可訓練變量a：',a)
a.assign([4,2,3])                       #   將可訓練變量改變
print('改變后的a：',a)
a.assign_add([4,0,5])                   #   將變量相加
print('相加后的變量a:',a)
a.assign_sub([8,8,8])                   #   將變量相減
print('相減后的變量a:',a)
print()
###############     isinstance()    ###############
a = tf.constant(5)
b = tf.Variable(5)
print('a:{}\nb{}'.format(a,b))
print("isinstance(a,tf.Tensor):{},isinstance(a,tf.Variable):{}".format(isinstance(a,tf.Tensor),isinstance(a,tf.Variable)))
print("isinstance(b,tf.Tensor):{},isinstance(b,tf.Variable):{}".format(isinstance(b,tf.Tensor),isinstance(b,tf.Variable)))