Question

我想在张量流中累积1到10。但这行不通。谁能帮我解决这个问题。

one = tf.constant(1)
value = tf.Variable(0,name="value")
increasing_op = tf.add(value,one)
assign_op = tf.assign(value,increasing_op)

#sum
sumvalue = tf.Variable(0,name = "sumvalue")
sum_op = tf.add(sumvalue,assign_op)

init = tf.global_variables_initializer()

with tf.Session() as session:
    session.run(init)
    for _ in range(10):
        session.run(sum_op)
    print(session.run(sumvalue))

Answer 1

sum_op返回计算后的值。您也可以利用tf.assign_add：

sumvalue = tf.Variable(0,name = "sumvalue")
sum_op = tf.assign_add(sumvalue, 1)

init = tf.global_variables_initializer()

with tf.Session() as session:
    session.run(init)
    for _ in range(10):
        sum_value = session.run(sum_op)
    print(sum_value)

Answer 2

感谢您的帮助。我修改了如下代码

model

是否可以将占位符用于n

如何在张量流中累积1到10？

2 个答案: