TensorFlow中的多线程

#encoding=utf-8
import threading
import numpy as np
import tensorflow as tf
#创建一个函数实现多线程，参数为Coordinater和线程号
def func(coord, t_id): 
count = 0 
while not coord.should_stop(): #不应该停止时计数 
print('thread ID:',t_id, 'count =', count)        
count += 1 
if(count == 5): #计到5时请求终止 
coord.request_stop() 
coord = tf.train.Coordinator()
 threads = [threading.Thread(target=func, args=(coord, i)) for i in range(4)]
#开始所有线程
for t in threads:    
t.start() 
coord.join(threads) #等待所有线程结束

#encoding=utf-8
import numpy as np
import tensorflow as tf batch_size = 2
#随机产生一个2*2的张量
example = tf.random_normal([2,2])
#创建一个RandomShuffleQueue，参数意义参见API
q = tf.RandomShuffleQueue(     capacity=1000,      min_after_dequeue=0,     dtypes=tf.float32,     shapes=[2,2])
#enqueue op，每次push一个张量
enq_op = q.enqueue(example)
#dequeue op, 每次取出batch_size个张量
xs = q.dequeue_many(batch_size)
#创建QueueRunner，包含4个enqueue op线程
qr = tf.train.QueueRunner(q, [enq_op]*4) coord = tf.train.Coordinator() sess = tf.Session()
#启动QueueRuner，开始线程
enq_threads = qr.create_threads(sess, coord=coord, start=True)for i in range(10):    
if coord.should_stop():        
break     print('step:', i, sess.run(xs)) 
#打印结果
coord.request_stop() coord.join(enq_threads)