tensorflow 多GPU编程 彻底指南
目前已有不少介绍tensorflow使用多GPU的文章，但大多凌乱不堪，更有相互借鉴之嫌。笔者钻研很多天，总算理清里面的脉络，特成此文以飨读者。python
- 缘起
tensorflow使用GPU时默认占满全部可用GPU的显存，但只在第一个GPU上进行计算。下图展现了一个典型的使用GPU训练的例子，虽然机器上有两块GPU，但却只有一块真正在工做，若是不加以利用，另外一块GPU就白白浪费了。咱们知道，GPU是一种相对比较昂贵的计算资源，虽然正值矿难，相比以前动辄八九千一块1080Ti的价格低了很多，但也不是通常人能浪费的起的，所以如何有效提升GPU特别是tensorflow上的利用率就成为了一项重要的考量。git
2.朴素的解决方案
一种常常说起的方法是设置可见的GPU，方法是经过设置CUDA_VISIBLE_DEVICES来完成，若是在shell中运行，每次运行某个shell脚本以前，加上export CUDA_VISIBLE_DEVICES=0#或者是你指望运行的GPU id（0到n-1,其中n是机器上GPU卡的数量），若是是在python脚本，还能够在全部代码以前:github
import os os.environ["CUDA_VISIBLE_DEVICES"]="0"
另外为缓解tensorflow一上来就占满整个GPU致使的问题，还能够在建立sess的时候传入附加的参数，更好的控制GPU的使用方法是shell
config = tf.ConfigProto(allow_soft_placement=True,allow_grouth=True) config.gpu_options.per_process_gpu_memory_fraction = 0.9 #占用90%显存 sess = tf.Session(config=config)
3.多GPU编程指南
上面的方法确实能够解决必定的问题，好比多个用户公用几块卡，每一个人分配不一样的卡来作实验，或者每张卡上运行不一样的参数设置。但有时候咱们更须要利用好已有的卡，来或得线性的加速比，以便在更短的时候获取参考结果，上面的方法就无能无力，只能本身写代码解决了，而当前怎么把单GPU代码转换为多GPU代码好像尚未一篇文章给出示例。编程
说到tensorflow的多GPU编程，不得不说是一部悲壮的血泪史。用过caffe的用户都知道caffe下的多GPU是多么的简单，从单GPU切换到多GPU你根本不用修改任何代码，只须要在编译选项中把USE_NCCL打开就好，剩下的都是caffe帮你自动完成了。到了tensorflow就不同的，全部的东西都得你本身操心，包括变量放哪，平均梯度什么的。固然好处是你对训练的过程理解的也更透彻了。网络
tensorflow的models里面有一个使用多GPU的例子，路径为tutorials/image/cifar10/cifar10_multi_gpu_train.py,但它仅仅是一个能运行的demo而已，里面并无解释枚举代码的含义，更别提为何写那些代码了，乱的一比。app
从网上找出一个能用且好用的例子并不像想象中的简单，我踏遍整个网络，把百度和谷歌中全部相关的文章都看了一遍，以为还不错的是Multi-GPU Basic，里面用一个小例子对比了单GPU和多GPU的差别。异步
另一个用Mnist的例子是这边文章要描述的重点，不过里面仍有一些干扰咱们理解的代码，好比那个assign_to_device函数。函数
其实tensorflow使用多GPU也是很是简单的，无外乎用tf.device("/gpu:0")之类的包起来，但要注意变量存放的位置。测试
咱们也以一个mnist单GPU训练的例子做为开始:
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
mnist=input_data.read_data_sets("/tmp/mnist/",one_hot=True)
num_gpus=2
num_steps=200
learning_rate=0.001
batch_size=1024
display_step=10
num_input=784
num_classes=10
def conv_net(x,is_training):
# "updates_collections": None is very import ,without will only get 0.10
batch_norm_params = {"is_training": is_training, "decay": 0.9, "updates_collections": None}
#,'variables_collections': [ tf.GraphKeys.TRAINABLE_VARIABLES ]
with slim.arg_scope([slim.conv2d, slim.fully_connected],
activation_fn=tf.nn.relu,
weights_initializer=tf.truncated_normal_initializer(mean=0.0, stddev=0.01),
weights_regularizer=slim.l2_regularizer(0.0005),
normalizer_fn=slim.batch_norm, normalizer_params=batch_norm_params):
with tf.variable_scope("ConvNet",reuse=tf.AUTO_REUSE):
x = tf.reshape(x, [-1, 28, 28, 1])
net = slim.conv2d(x, 6, [5,5], scope="conv_1")
net = slim.max_pool2d(net, [2, 2],scope="pool_1")
net = slim.conv2d(net, 12, [5,5], scope="conv_2")
net = slim.max_pool2d(net, [2, 2], scope="pool_2")
net = slim.flatten(net, scope="flatten")
net = slim.fully_connected(net, 100, scope="fc")
net = slim.dropout(net,is_training=is_training)
net = slim.fully_connected(net, num_classes, scope="prob", activation_fn=None,normalizer_fn=None)
return net
def train_single():
X = tf.placeholder(tf.float32, [None, num_input])
Y = tf.placeholder(tf.float32, [None, num_classes])
logits=conv_net(X,True)
loss=tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=Y,logits=logits))
opt=tf.train.AdamOptimizer(learning_rate)
train_op=opt.minimize(loss)
logits_test=conv_net(X,False)
correct_prediction = tf.equal(tf.argmax(logits_test, 1), tf.argmax(Y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for step in range(1,num_steps+1):
batch_x, batch_y = mnist.train.next_batch(batch_size)
sess.run(train_op,feed_dict={X:batch_x,Y:batch_y})
if step%display_step==0 or step==1:
loss_value,acc=sess.run([loss,accuracy],feed_dict={X:batch_x,Y:batch_y})
print("Step:" + str(step) + ":" + str(loss_value) + " " + str(acc))
print("Done")
print("Testing Accuracy:",np.mean([sess.run(accuracy, feed_dict={X: mnist.test.images[i:i + batch_size],
Y: mnist.test.labels[i:i + batch_size]}) for i in
range(0, len(mnist.test.images), batch_size)]))
if __name__ == "__main__":
train_single()
在上面寥寥数行的代码中，咱们下载并构建了mnist数据集，经过mnist.train.next_batch(batch_szie)的方式返回一个batch的数据，而后扔进conv_net网络，计算出loss后再使用Adam优化器进行训练，最后测试了其精度。
只须要几分钟，在个人一块1080Ti卡上就能运行完毕，而且获得至关好的结果:
Step:1:2.213318 0.20703125
Step:10:0.46338645 0.88183594
Step:20:0.18729115 0.9550781
Step:30:0.17860937 0.9628906
Step:40:0.11540267 0.97558594
Step:50:0.081396215 0.9824219
Step:60:0.097750194 0.9746094
Step:70:0.060169913 0.984375
Step:80:0.059070613 0.98828125
Step:90:0.060746174 0.9892578
Step:100:0.057775088 0.9892578
Step:110:0.038614694 0.98828125
Step:120:0.0369242 0.9921875
Step:130:0.035249908 0.9941406
Step:140:0.03395287 0.9902344
Step:150:0.03798459 0.98828125
Step:160:0.052775905 0.99121094
Step:170:0.017296169 0.99609375
Step:180:0.026407585 0.9951172
Step:190:0.044104658 0.9941406
Step:200:0.025472593 0.99121094
Done
('Testing Accuracy:', 0.99301463)
接下来咱们将以此为起点，对其进行多GPU化改造。下面的图很好的阐释了整个流程。
多GPU并行可分为模型并行和数据并行两大类，上图展现的是数据并行，这也是咱们常常用到的方法，而其中数据并行又可分为同步方式和异步方式两种，因为咱们通常都会配置一样的显卡，所以这儿也选择了同步方式，也就是把数据分给不一样的卡，等全部的GPU都计算完梯度后进行平均，最后再更新梯度。
首先要改造的就是数据读取部分，因为如今咱们有多快卡，每张卡要分到不一样的数据，因此在获取batch的时候要把大小改成batch_x,batch_y=mnist.train.next_batch(batch_size*num_gpus)，一次取足够的数据保证每块卡都分到batch_size大小的数据。而后咱们对取到的数据进行切分，咱们以i表示GPU的索引，连续的batch_size大小的数据分给同一块GPU:
_x=X[i*batch_size:(i+1)*batch_size] _y=Y[i*batch_size:(i+1)*batch_size]
因为咱们多个GPU上共享一样的图，为了防止名字混乱，最好使用name_scope进行区分，也就是以下的形式:
for i in range(2):
with tf.device("/gpu:%d"%i):
with tf.name_scope("tower_%d"%i):
_x=X[i*batch_size:(i+1)*batch_size]
_y=Y[i*batch_size:(i+1)*batch_size]
logits=conv_net(_x,dropout,reuse_vars,True)
咱们须要有个列表存储全部GPU上的梯度，还有就是复用变量，须要在以前定义以下两个值:
tower_grads=[] reuse_vars=False
全部的准备工做都已完成，就能够计算每一个GPU上的梯度了
opt = tf.train.AdamOptimizer(learning_rate)
loss=tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=_y,logits=logits))
grads=opt.compute_gradients(loss)
reuse_vars=True
tower_grads.append(grads)
这样tower_grads就存储了全部GPU上全部变量的梯度，下面就是计算平均值了，这个是全部见过的函数中惟一一个几乎从没变过的代码:
def average_gradients(tower_grads):
average_grads=[]
for grad_and_vars in zip(*tower_grads):
grads=[]
for g,_ in grad_and_vars:
expend_g=tf.expand_dims(g,0)
grads.append(expend_g)
grad=tf.concat(grads,0)
grad=tf.reduce_mean(grad,0)
v=grad_and_vars[0][1]
grad_and_var=(grad,v)
average_grads.append(grad_and_var)
return average_grads
tower_grads里面保存的形式是(第一个GPU上的梯度，第二个GPU上的梯度,...第N-1个GPU上的梯度)，这里有一点须要注意的是zip(*)，它的做用上把上面的那个列表转换成((grad0_gpu0, var0_gpu0), ... , (grad0_gpuN, var0_gpuN))的形式，也就是以列访问的方式，取到的就是某个变量在不一样GPU上的值。
最后就是更新梯度了
grads=average_gradients(tower_grads) train_op=opt.apply_gradients(grads)
上面的讲述略有零散，最后咱们给个全代码版本方便你们测试:
import tensorflow as tf
import numpy as np
from tensorflow.contrib import slim
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/tmp/mnist/", one_hot=True)
num_gpus = 2
num_steps = 1000
learning_rate = 0.001
batch_size = 1000
display_step = 10
num_input = 784
num_classes = 10
def conv_net_with_layers(x,is_training,dropout = 0.75):
with tf.variable_scope("ConvNet", reuse=tf.AUTO_REUSE):
x = tf.reshape(x, [-1, 28, 28, 1])
x = tf.layers.conv2d(x, 12, 5, activation=tf.nn.relu)
x = tf.layers.max_pooling2d(x, 2, 2)
x = tf.layers.conv2d(x, 24, 3, activation=tf.nn.relu)
x = tf.layers.max_pooling2d(x, 2, 2)
x = tf.layers.flatten(x)
x = tf.layers.dense(x, 100)
x = tf.layers.dropout(x, rate=dropout, training=is_training)
out = tf.layers.dense(x, 10)
out = tf.nn.softmax(out) if not is_training else out
return out
def conv_net(x,is_training):
# "updates_collections": None is very import ,without will only get 0.10
batch_norm_params = {"is_training": is_training, "decay": 0.9, "updates_collections": None}
#,'variables_collections': [ tf.GraphKeys.TRAINABLE_VARIABLES ]
with slim.arg_scope([slim.conv2d, slim.fully_connected],
activation_fn=tf.nn.relu,
weights_initializer=tf.truncated_normal_initializer(mean=0.0, stddev=0.01),
weights_regularizer=slim.l2_regularizer(0.0005),
normalizer_fn=slim.batch_norm, normalizer_params=batch_norm_params):
with tf.variable_scope("ConvNet",reuse=tf.AUTO_REUSE):
x = tf.reshape(x, [-1, 28, 28, 1])
net = slim.conv2d(x, 6, [5,5], scope="conv_1")
net = slim.max_pool2d(net, [2, 2],scope="pool_1")
net = slim.conv2d(net, 12, [5,5], scope="conv_2")
net = slim.max_pool2d(net, [2, 2], scope="pool_2")
net = slim.flatten(net, scope="flatten")
net = slim.fully_connected(net, 100, scope="fc")
net = slim.dropout(net,is_training=is_training)
net = slim.fully_connected(net, num_classes, scope="prob", activation_fn=None,normalizer_fn=None)
return net
def average_gradients(tower_grads):
average_grads = []
for grad_and_vars in zip(*tower_grads):
grads = []
for g, _ in grad_and_vars:
expend_g = tf.expand_dims(g, 0)
grads.append(expend_g)
grad = tf.concat(grads, 0)
grad = tf.reduce_mean(grad, 0)
v = grad_and_vars[0][1]
grad_and_var = (grad, v)
average_grads.append(grad_and_var)
return average_grads
def train():
with tf.device("/cpu:0"):
global_step=tf.train.get_or_create_global_step()
tower_grads = []
X = tf.placeholder(tf.float32, [None, num_input])
Y = tf.placeholder(tf.float32, [None, num_classes])
opt = tf.train.AdamOptimizer(learning_rate)
with tf.variable_scope(tf.get_variable_scope()):
for i in range(2):
with tf.device("/gpu:%d" % i):
with tf.name_scope("tower_%d" % i):
_x = X[i * batch_size:(i + 1) * batch_size]
_y = Y[i * batch_size:(i + 1) * batch_size]
logits = conv_net(_x, True)
tf.get_variable_scope().reuse_variables()
loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=_y, logits=logits))
grads = opt.compute_gradients(loss)
tower_grads.append(grads)
if i == 0:
logits_test = conv_net(_x, False)
correct_prediction = tf.equal(tf.argmax(logits_test, 1), tf.argmax(_y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
grads = average_gradients(tower_grads)
train_op = opt.apply_gradients(grads)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for step in range(1, num_steps + 1):
batch_x, batch_y = mnist.train.next_batch(batch_size * num_gpus)
sess.run(train_op, feed_dict={X: batch_x, Y: batch_y})
if step % 10 == 0 or step == 1:
loss_value, acc = sess.run([loss, accuracy], feed_dict={X: batch_x, Y: batch_y})
print("Step:" + str(step) + ":" + str(loss_value) + " " + str(acc))
print("Done")
print("Testing Accuracy:",
np.mean([sess.run(accuracy, feed_dict={X: mnist.test.images[i:i + batch_size],
Y: mnist.test.labels[i:i + batch_size]}) for i in
range(0, len(mnist.test.images), batch_size)]))
def train_single():
X = tf.placeholder(tf.float32, [None, num_input])
Y = tf.placeholder(tf.float32, [None, num_classes])
logits=conv_net(X,True)
loss=tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=Y,logits=logits))
opt=tf.train.AdamOptimizer(learning_rate)
train_op=opt.minimize(loss)
logits_test=conv_net(X,False)
correct_prediction = tf.equal(tf.argmax(logits_test, 1), tf.argmax(Y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for step in range(1,num_steps+1):
batch_x, batch_y = mnist.train.next_batch(batch_size)
sess.run(train_op,feed_dict={X:batch_x,Y:batch_y})
if step%display_step==0 or step==1:
loss_value,acc=sess.run([loss,accuracy],feed_dict={X:batch_x,Y:batch_y})
print("Step:" + str(step) + ":" + str(loss_value) + " " + str(acc))
print("Done")
print("Testing Accuracy:",np.mean([sess.run(accuracy, feed_dict={X: mnist.test.images[i:i + batch_size],
Y: mnist.test.labels[i:i + batch_size]}) for i in
range(0, len(mnist.test.images), batch_size)]))
if __name__ == "__main__":
#train_single()
train()
我这边的运行结果是
Step:1:2.1529438 0.27929688
Step:10:0.4463266 0.89941406
Step:20:0.18885617 0.9580078
Step:30:0.12153786 0.96484375
Step:40:0.07257775 0.98339844
Step:50:0.07452829 0.98535156
Step:60:0.048265547 0.99121094
Step:70:0.02948389 0.9873047
Step:80:0.048876762 0.9902344
Step:90:0.06494201 0.9902344
Step:100:0.024681691 0.98535156
Step:110:0.025596365 0.99316406
Step:120:0.02779768 0.9941406
Step:130:0.02191917 0.9980469
Step:140:0.022235561 0.9951172
Step:150:0.0124597605 0.9970703
Step:160:0.011557209 0.9902344
Step:170:0.014788041 0.99609375
Step:180:0.015063373 0.99609375
Step:190:0.012394376 0.99902344
Step:200:0.017424839 0.99609375
Done
('Testing Accuracy:', 0.99301463)
能够看到GPU都用上了