手写数字识别，神经网络交叉商结果正确，正确率总是不变

Question

我遇到的问题是，经过训练，测试集的交叉商下降非常快，但是正确率总是不变，真是想不到为什么
#!/usr/bin/env python3

-*- coding: utf-8 -*-

import tensorflow as tf
import numpy as np
from tensorflow.contrib.layers import fully_connected
from tensorflow.examples.tutorials.mnist import input_data

x = tf.placeholder(dtype=tf.float32,shape=[None,784])
y = tf.placeholder(dtype=tf.float32,shape=[None,10])
test_x = tf.placeholder(dtype=tf.float32,shape=[None,784])
test_y = tf.placeholder(dtype=tf.float32,shape=[None,10])

mnist = input_data.read_data_sets("/home/xuenzhu/mnist_data", one_hot=True)

hidden1 = fully_connected(x,100,activation_fn=tf.nn.relu,
weights_initializer=tf.random_normal_initializer())

hidden2 = fully_connected(hidden1,100,activation_fn=tf.nn.relu,
weights_initializer=tf.random_normal_initializer())

outputs = fully_connected(hidden2,10,activation_fn=tf.nn.relu,
weights_initializer=tf.random_normal_initializer())

loss = tf.nn.softmax_cross_entropy_with_logits(labels=y,logits=outputs)
reduce_mean_loss = tf.reduce_mean(loss)

equal_result = tf.equal(tf.argmax(outputs,1),tf.argmax(y,1))
cast_result = tf.cast(equal_result,dtype=tf.float32)
accuracy = tf.reduce_mean(cast_result)

train_op = tf.train.AdamOptimizer(0.001).minimize(reduce_mean_loss)

with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
print(sess.run(accuracy,feed_dict={x:mnist.test.images,y:mnist.test.labels}))
for i in range(10000):
xs,ys = mnist.train.next_batch(100)
sess.run(train_op,feed_dict={x:xs,y:ys})
if i%100==0:
print(sess.run(reduce_mean_loss,feed_dict={x:mnist.test.images,y:mnist.test.labels}))
print(sess.run(accuracy,feed_dict={x:mnist.test.images,y:mnist.test.labels}))

Answer 1

正确率是多少？如果保持在0.1左右，说明没有学习到（因为有10个分类，随机的权重识别出来正确概率就是0.1）
Adam换成SGD，学习率设置小一点看看。调试下，输出下损失函数的损失率。