#coding:utf-8

import  tensorflow as tf

from  tensorflow.examples.tutorials.mnist import  input_data

### 注意

# init_state = tf.zeros(shape=[batch_size,rnn_cell.state_size])

# init_state = lstm_cell.zero_state(batch_size, dtype=tf.float32)

mnist=input_data.read_data_sets("./data",one_hot=True)

# 常规参数

train_rate=0.001

train_step=10000

batch_size=1280

display_step=100

# rnn参数

frame_size=28           # 输入特征数

sequence_length=28      # 输入个数, 时序

hidden_num=100          # 隐层神经元个数

n_classes=10

# 定义输入,输出

# 此处输入格式是样本数*特征数，特征是把图片拉成一维的，当然一维还是二维自己定，改成相应的代码就行了

x=tf.placeholder(dtype=tf.float32,shape=[None,sequence_length*frame_size],name="inputx")

y=tf.placeholder(dtype=tf.float32,shape=[None,n_classes],name="expected_y")

# 定义权值

# 注意权值设定只设定v， u和w无需设定

weights=tf.Variable(tf.truncated_normal(shape=[hidden_num,n_classes]))      # 全连接层权重

bias=tf.Variable(tf.zeros(shape=[n_classes]))

def RNN(x,weights,bias):

    x=tf.reshape(x,shape=[-1,sequence_length,frame_size])   # 3维

    rnn_cell=tf.nn.rnn_cell.BasicRNNCell(hidden_num)

    ### 注意

    # init_state=tf.zeros(shape=[batch_size,rnn_cell.state_size])         # rnn_cell.state_size 100

    init_state=rnn_cell.zero_state(batch_size, dtype=tf.float32)

    output,states=tf.nn.dynamic_rnn(rnn_cell,x,initial_state=init_state,dtype=tf.float32)

    return tf.nn.softmax(tf.matmul(output[:,-1,:],weights)+bias,1)  # y=softmax(vh+c)

predy=RNN(x,weights,bias)

# 以下所有神经网络大同小异

cost=tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=predy,labels=y))

train=tf.train.AdamOptimizer(train_rate).minimize(cost)

correct_pred=tf.equal(tf.argmax(predy,1),tf.argmax(y,1))

accuracy=tf.reduce_mean(tf.to_float(correct_pred))

sess=tf.Session()

sess.run(tf.global_variables_initializer())

step=1

testx,testy=mnist.test.next_batch(batch_size)

while step<train_step:

    batch_x,batch_y=mnist.train.next_batch(batch_size)

    _loss,__=sess.run([cost,train],feed_dict={x:batch_x,y:batch_y})

    if step % display_step ==0:

        print()

        acc,loss=sess.run([accuracy,cost],feed_dict={x:testx,y:testy})

        print(step,acc,loss)

    step+=1

init_state=tf.zeros(shape=[batch_size,rnn_cell.state_size])         # rnn_cell.state_size 100，100为节点数

init_state=rnn_cell.zero_state(batch_size, dtype=tf.float32)