使用 GPUs
支持的设备
在一套标准的系统上通常有多个计算设备. TensorFlow 支持 CPU 和 GPU 这两种设备. 我们用指定字符串strings
来标识这些设备. 比如:
-
"/cpu:0"
: 机器中的 CPU -
"/gpu:0"
: 机器中的 GPU, 如果你有一个的话. -
"/gpu:1"
: 机器中的第二个 GPU, 以此类推...
如果一个 TensorFlow 的 operation 中兼有 CPU 和 GPU 的实现, 当这个算子被指派设备时, GPU 有优先权. 比如matmul
中 CPU和 GPU kernel 函数都存在. 那么在cpu:0
和gpu:0
中, matmul
operation 会被指派给 gpu:0
.
记录设备指派情况
为了获取你的 operations 和 Tensor 被指派到哪个设备上运行, 用 log_device_placement
新建一个session
, 并设置为True
.
# 新建一个 graph.
a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a')
b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b')
c = tf.matmul(a, b)
# 新建session with log_device_placement并设置为True.
sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
# 运行这个 op.
print sess.run(c)
你应该能看见以下输出:
Device mapping:
/job:localhost/replica:0/task:0/gpu:0 -> device: 0, name: Tesla K40c, pci bus
id: 0000:05:00.0
b: /job:localhost/replica:0/task:0/gpu:0
a: /job:localhost/replica:0/task:0/gpu:0
MatMul: /job:localhost/replica:0/task:0/gpu:0
[[ 22. 28.]
[ 49. 64.]]
手工指派设备
如果你不想使用系统来为 operation 指派设备, 而是手工指派设备, 你可以用 with tf.device
创建一个设备环境, 这个环境下的 operation 都统一运行在环境指定的设备上.
# 新建一个graph.
with tf.device('/cpu:0'):
a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a')
b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b')
c = tf.matmul(a, b)
# 新建session with log_device_placement并设置为True.
sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
# 运行这个op.
print sess.run(c)
你会发现现在 a
和 b
操作都被指派给了 cpu:0
.
Device mapping:
/job:localhost/replica:0/task:0/gpu:0 -> device: 0, name: Tesla K40c, pci bus
id: 0000:05:00.0
b: /job:localhost/replica:0/task:0/cpu:0
a: /job:localhost/replica:0/task:0/cpu:0
MatMul: /job:localhost/replica:0/task:0/gpu:0
[[ 22. 28.]
[ 49. 64.]]
在多GPU系统里使用单一GPU
如果你的系统里有多个 GPU, 那么 ID 最小的 GPU 会默认使用. 如果你想用别的 GPU, 可以用下面的方法显式的声明你的偏好:
# 新建一个 graph.
with tf.device('/gpu:2'):
a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a')
b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b')
c = tf.matmul(a, b)
# 新建 session with log_device_placement 并设置为 True.
sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
# 运行这个 op.
print sess.run(c)
如果你指定的设备不存在, 你会收到 InvalidArgumentError
错误提示:
InvalidArgumentError: Invalid argument: Cannot assign a device to node 'b':
Could not satisfy explicit device specification '/gpu:2'
[[Node: b = Const[dtype=DT_FLOAT, value=Tensor<type: float shape: [3,2]
values: 1 2 3...>, _device="/gpu:2"]()]]
为了避免出现你指定的设备不存在这种情况, 你可以在创建的 session
里把参数 allow_soft_placement
设置为True
, 这样 tensorFlow 会自动选择一个存在并且支持的设备来运行 operation.
# 新建一个 graph.
with tf.device('/gpu:2'):
a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a')
b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b')
c = tf.matmul(a, b)
# 新建 session with log_device_placement 并设置为 True.
sess = tf.Session(config=tf.ConfigProto(
allow_soft_placement=True, log_device_placement=True))
# 运行这个 op.
print sess.run(c)
使用多个 GPU
如果你想让 TensorFlow 在多个 GPU 上运行, 你可以建立 multi-tower 结构, 在这个结构里每个 tower 分别被指配给不同的 GPU 运行. 比如:
# 新建一个 graph.
c = []
for d in ['/gpu:2', '/gpu:3']:
with tf.device(d):
a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3])
b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2])
c.append(tf.matmul(a, b))
with tf.device('/cpu:0'):
sum = tf.add_n(c)
# 新建session with log_device_placement并设置为True.
sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
# 运行这个op.
print sess.run(sum)
你会看到如下输出:
Device mapping:
/job:localhost/replica:0/task:0/gpu:0 -> device: 0, name: Tesla K20m, pci bus
id: 0000:02:00.0
/job:localhost/replica:0/task:0/gpu:1 -> device: 1, name: Tesla K20m, pci bus
id: 0000:03:00.0
/job:localhost/replica:0/task:0/gpu:2 -> device: 2, name: Tesla K20m, pci bus
id: 0000:83:00.0
/job:localhost/replica:0/task:0/gpu:3 -> device: 3, name: Tesla K20m, pci bus
id: 0000:84:00.0
Const_3: /job:localhost/replica:0/task:0/gpu:3
Const_2: /job:localhost/replica:0/task:0/gpu:3
MatMul_1: /job:localhost/replica:0/task:0/gpu:3
Const_1: /job:localhost/replica:0/task:0/gpu:2
Const: /job:localhost/replica:0/task:0/gpu:2
MatMul: /job:localhost/replica:0/task:0/gpu:2
AddN: /job:localhost/replica:0/task:0/cpu:0
[[ 44. 56.]
[ 98. 128.]]
cifar10 tutorial 这个例子很好的演示了怎样用GPU集群训练.
==============programe with GPU==================
下面是带GPU的Minist程序代码:
# encoding=utf8
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from tensorflow.examples.tutorials.mnist import input_data
import tensorflow as tf
flags = tf.app.flags
FLAGS = flags.FLAGS
flags.DEFINE_string('data_dir', '/tmp/data/', 'Directory for storing data')
global mnist
mnist = input_data.read_data_sets('/home/yuan/testMinist', one_hot=True)
# mnist = input_data.read_data_sets('/home/yuan/Xia', one_hot=True)
#远程下载MNIST数据,建议先下载好并保存在MNIST_data目录下
def DownloadData():
global mnist
#mnist = input_data.read_data_sets('/home/yuan/testMinist', one_hot=True) #编码格式:one-hot
#print(mnist.train.images.shape, mnist.train.labels.shape)
#print(mnist.test.images.shape, mnist.test.labels.shape)
#print(mnist.validation.images.shape, mnist.validation.labels.shape)
def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev=0.1)
return tf.Variable(initial)
def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)
def conv2d(x, W):
return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')
def max_pool_2x2(x):
return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
def Train():
with tf.device('/cpu:0'):
x = tf.placeholder("float", shape=[None, 784]) #构建占位符,代表输入的图像,None表示样本的数量可以是任意的
W = tf.Variable(tf.zeros([784,10])) #构建一个变量,代表训练目标weights,初始化为0
b = tf.Variable(tf.zeros([10]))
y_ = tf.placeholder("float", shape=[None, 10])
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.333)
#sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True, log_device_placement=True))
#sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
#sess = tf.InteractiveSession(config=tf.ConfigProto(gpu_options=gpu_options))
sess = tf.InteractiveSession(config=config)
#sess = tf.Session(config=config)
with tf.device('/gpu:0'):
#with tf.device('/cpu:0'):
#sess = tf.InteractiveSession()
#Step 1
#定义算法公式Softmax Regression
# = tf.placeholder("float", shape=[None, 784]) #构建占位符,代表输入的图像,None表示样本的数量可以是任意的
# = tf.Variable(tf.zeros([784,10])) #构建一个变量,代表训练目标weights,初始化为0
#b = tf.Variable(tf.zeros([10])) #构建一个变量,代表训练目标biases,初始化为0
#y = tf.nn.softmax(tf.matmul(x, W) + b) #构建了一个softmax的模型:y = softmax(Wx + b),y指样本标签的预测值
W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])
x_image = tf.reshape(x, [-1, 28, 28, 1])
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)
W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])
h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)
# Now image size is reduced to 7*7
W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])
h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)
keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])
y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
#Step 2
#定义损失函数,选定优化器,并指定优化器优化损失函数
#y_ = tf.placeholder(tf.float32, [None, 10]) # 构建占位符,代表样本标签的真实值
#y_ = tf.placeholder("float", shape=[None, 10])
# 交叉熵损失函数
with tf.device('/cpu:0'):
cross_entropy = -tf.reduce_sum(y_ * tf.log(y_conv))
#y = tf.matmul(x, W) + b
#cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y))
# 使用梯度下降法(0.01的学习率)来最小化这个交叉熵损失函数
#train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
#Step 3
#使用随机梯度下降训练数据
#tf.global_variables_initializer().run() #when the comple enviroment is python3, changing global_variables_initializer
#tf.initialize_all_variables().run() #when the comple enviroment is python2, changing initialize_all_variables()
#for i in range(10000): #迭代次数为1000
# batch_xs, batch_ys = mnist.train.next_batch(100) #使用minibatch的训练数据,一个batch的大小为100
# sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys}) #用训练数据替代占位符来执行训练
#Step 4
#在测试集上对准确率进行评测
correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1)) #tf.argmax()返回的是某一维度上其数据最大所在的索引值,在这里即代表预测值和真值
#accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32)) #用平均值来统计测试准确率
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float")) #用平均值来统计测试准确率
#sess.run(tf.initialize_all_variables()) # when compile in python2,changing this.
sess.run(tf.global_variables_initializer()) # when compile in python3,changing this.
#print(sess.run(accuracy, feed_dict={x: mnist.test.images, y_: mnist.test.labels})) #打印测试信息
for i in range(2000): #迭代次数为1000
#batch_xs, batch_ys = mnist.train.next_batch(50) #使用minibatch的训练数据,一个batch的大小为100
batch = mnist.train.next_batch(50)
#sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys}) #用训练数据替代占位符来执行训练
if i%100 == 0: # this step is conducted when the condition needs
train_accuracy = accuracy.eval(feed_dict={
x:batch[0], y_: batch[1], keep_prob: 1.0})
print ("step %d, training accuracy %.3f" %(i, train_accuracy)) # the output the training accuracy when the condition needs
train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5}) # this step is used to get an trained model
#sess.close()
print ("Training finished")
print ("test accuracy %.3f" % accuracy.eval(feed_dict={
x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))
if __name__ == '__main__':
DownloadData();
Train();
==============programe without GPU, only using CPU==================
下面是带CPU的Minist程序代码:
#!/usr/bin/python
# -*- coding: utf-8 -*-
import input_data
mnist = input_data.read_data_sets('/home/yuan/testMinist', one_hot=True)
import tensorflow as tf
import sys
#from tensorflow.examples.tutorials.mnist import input_data
from tensorflow.examples.tutorials.mnist import input_data
#print("step",mnist)
#print("step",mnist.train)
#print("step",mnist)
def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev=0.1)
return tf.Variable(initial)
def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)
def conv2d(x, W):
return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')
def max_pool_2x2(x):
return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
#mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)
sess = tf.InteractiveSession()
x = tf.placeholder("float", shape=[None, 784])
y_ = tf.placeholder("float", shape=[None, 10])
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))
W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])
x_image = tf.reshape(x, [-1, 28, 28, 1])
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)
W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])
h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)
# Now image size is reduced to 7*7
W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])
h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)
keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])
y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
cross_entropy = -tf.reduce_sum(y_*tf.log(y_conv))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))
sess.run(tf.initialize_all_variables())
for i in range(2000): # the iterate steps for array
batch = mnist.train.next_batch(50)
#print("the format shape is:",batch[0].shape)
#print("the format size is:",batch[0].size)
#print("the format 2 is:",batch[0])
#print("the format 2 is:",batch[0].length)
#print("the format 2 is:",batch.size)
if i%100 == 0: # this step is conducted when the condition needs
train_accuracy = accuracy.eval(feed_dict={
x:batch[0], y_: batch[1], keep_prob: 1.0})
print "step %d, training accuracy %.3f"%(i, train_accuracy) # the output the training accuracy when the condition needs
train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5}) # this step is used to get an trained model
print "Training finished"
print "test accuracy %.3f" % accuracy.eval(feed_dict={
x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0})