用Keras搭建神经网络 简单模版（五）——RNN LSTM Regressor 循环神经网络

# -*- coding: utf-8 -*-
import numpy as np
np.random.seed(1337)
import matplotlib.pyplot as plt
from keras.models import Sequential
from keras.layers import LSTM,TimeDistributed,Dense
from keras.optimizers import Adam

BATCH_START = 0 
TIME_STEPS = 20 
BATCH_SIZE = 50
INPUT_SIZE = 1 
OUTPUT_SIZE = 1
CELL_SIZE = 20
LR = 0.006

def get_batch():
    global BATCH_START,TIME_STEPS
    # xs shape(50,20,)
    #xs=np.arange(0,0+20*50).reshape(50,20)
    xs = np.arange(BATCH_START,BATCH_START+TIME_STEPS*BATCH_SIZE).reshape((BATCH_SIZE,TIME_STEPS)) / (10*np.pi)
    seq = np.sin(xs)
    res = np.cos(xs)
    BATCH_START += TIME_STEPS
    #plt.plot(xs[0,:],res[0,:],'r',xs[0,:],seq[0,:],'b--')
    #plt.show()
    return [seq[:,:,np.newaxis], res[:,:,np.newaxis],xs]

#get_batch()
#exit()
    
    
model = Sequential()

model.add(LSTM(output_dim=CELL_SIZE, 
               return_sequences=True, # 每一个时间点都输出一个output
               batch_input_shape=(BATCH_SIZE,TIME_STEPS,INPUT_SIZE),
               stateful = True,# batch和batch之间是否有联系
               # 前一个batch的最后一步和后一个batch的第一步是有联系的
        )) 

model.add(TimeDistributed(Dense(OUTPUT_SIZE))) # dense对每一个output连接，对每一个时间点都要计算

adam = Adam(LR)
model.compile(optimizer = adam,
              loss = 'mse',)

print('Training ------------')
for step in range(501):
    # data shape = (batch_num,steps,inputs/output)
    X_batch, Y_batch, xs = get_batch()
    cost = model.train_on_batch(X_batch, Y_batch)
    pred = model.predict(X_batch,BATCH_SIZE)
    plt.plot(xs[0,:], Y_batch[0].flatten(),'r',xs[0,:],pred.flatten()[:TIME_STEPS],'b--')
    plt.ylim((-1.2,1.2))
    plt.draw()
    plt.pause(0.5)
    if step % 10 == 0:
        print('train cost',cost)