Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
PaddlePaddle
book
提交
0fb04454
B
book
项目概览
PaddlePaddle
/
book
通知
16
Star
4
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
40
列表
看板
标记
里程碑
合并请求
37
Wiki
5
Wiki
分析
仓库
DevOps
项目成员
Pages
B
book
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
40
Issue
40
列表
看板
标记
里程碑
合并请求
37
合并请求
37
Pages
分析
分析
仓库分析
DevOps
Wiki
5
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
未验证
提交
0fb04454
编写于
6月 11, 2018
作者:
D
daminglu
提交者:
GitHub
6月 11, 2018
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Translate to Chinese: Chap 1 (#544)
上级
e5e810c3
变更
3
隐藏空白更改
内联
并排
Showing
3 changed file
with
95 addition
and
110 deletion
+95
-110
01.fit_a_line/README.cn.md
01.fit_a_line/README.cn.md
+93
-108
01.fit_a_line/README.md
01.fit_a_line/README.md
+1
-1
01.fit_a_line/train.py
01.fit_a_line/train.py
+1
-1
未找到文件。
01.fit_a_line/README.cn.md
浏览文件 @
0fb04454
...
...
@@ -47,22 +47,6 @@ $$MSE=\frac{1}{n}\sum_{i=1}^{n}{(\hat{Y_i}-Y_i)}^2$$
## 数据集
### 数据集接口的封装
首先加载需要的包
```
python
import
paddle.v2
as
paddle
import
paddle.v2.dataset.uci_housing
as
uci_housing
```
我们通过uci_housing模块引入了数据集合
[
UCI Housing Data Set
](
https://archive.ics.uci.edu/ml/datasets/Housing
)
其中,在uci_housing模块中封装了:
1.
数据下载的过程。下载数据保存在~/.cache/paddle/dataset/uci_housing/housing.data。
2.
[
数据预处理
](
#数据预处理
)
的过程。
### 数据集介绍
这份数据集共506行,每行包含了波士顿郊区的一类房屋的相关信息及该类房屋价格的中位数。其各维属性的意义如下:
...
...
@@ -110,157 +94,158 @@ import paddle.v2.dataset.uci_housing as uci_housing
`fit_a_line/trainer.py`
演示了训练的整体过程。
###
初始化PaddlePaddle
###
配置数据提供器(Datafeeder)
首先我们引入必要的库:
```
python
paddle
.
init
(
use_gpu
=
False
,
trainer_count
=
1
)
import
paddle
import
paddle.fluid
as
fluid
import
numpy
```
### 模型配置
我们通过uci_housing模块引入了数据集合
[
UCI Housing Data Set
](
https://archive.ics.uci.edu/ml/datasets/Housing
)
线性回归的模型其实就是一个采用线性激活函数(linear activation,
`LinearActivation`
)的全连接层(fully-connected layer,
`fc_layer`
)
:
其中,在uci_housing模块中封装了
:
```
python
x
=
paddle
.
layer
.
data
(
name
=
'x'
,
type
=
paddle
.
data_type
.
dense_vector
(
13
))
y_predict
=
paddle
.
layer
.
fc
(
input
=
x
,
size
=
1
,
act
=
paddle
.
activation
.
Linear
())
y
=
paddle
.
layer
.
data
(
name
=
'y'
,
type
=
paddle
.
data_type
.
dense_vector
(
1
))
cost
=
paddle
.
layer
.
square_error_cost
(
input
=
y_predict
,
label
=
y
)
```
1.
数据下载的过程。下载数据保存在~/.cache/paddle/dataset/uci_housing/housing.data。
2.
[
数据预处理
](
#数据预处理
)
的过程。
### 保存网络拓扑
接下来我们定义了用于训练和测试的数据提供器。提供器每次读入一个大小为
`BATCH_SIZE`
的数据批次。如果用户希望加一些随机性,她可以同时定义一个批次大小和一个缓存大小。这样的话,每次数据提供器会从缓存中随机读取批次大小那么多的数据。
```
python
# Save the inference topology to protobuf.
inference_topology
=
paddle
.
topology
.
Topology
(
layers
=
y_predict
)
with
open
(
"inference_topology.pkl"
,
'wb'
)
as
f
:
inference_topology
.
serialize_for_inference
(
f
)
```
BATCH_SIZE
=
20
### 创建参数
train_reader
=
paddle
.
batch
(
paddle
.
reader
.
shuffle
(
paddle
.
dataset
.
uci_housing
.
train
(),
buf_size
=
500
),
batch_size
=
BATCH_SIZE
)
```
python
parameters
=
paddle
.
parameters
.
create
(
cost
)
test_reader
=
paddle
.
batch
(
paddle
.
reader
.
shuffle
(
paddle
.
dataset
.
uci_housing
.
test
(),
buf_size
=
500
),
batch_size
=
BATCH_SIZE
)
```
### 创建Trainer
### 配置训练程序
训练程序的目的是定义一个训练模型的网络结构.对于线性回归来讲,它就是一个从输入到输出的简单的全连接层。更加复杂的结果,比如卷积神经网络,递归神经网络等会在随后的章节中介绍。训练程序必须返回
`平均损失`
作为第一个返回值,因为它会被后面反向传播算法所用到。
```
python
optimizer
=
paddle
.
optimizer
.
Momentum
(
momentum
=
0
)
def
train_program
():
y
=
fluid
.
layers
.
data
(
name
=
'y'
,
shape
=
[
1
],
dtype
=
'float32'
)
# feature vector of length 13
x
=
fluid
.
layers
.
data
(
name
=
'x'
,
shape
=
[
13
],
dtype
=
'float32'
)
y_predict
=
fluid
.
layers
.
fc
(
input
=
x
,
size
=
1
,
act
=
None
)
loss
=
fluid
.
layers
.
square_error_cost
(
input
=
y_predict
,
label
=
y
)
avg_loss
=
fluid
.
layers
.
mean
(
loss
)
trainer
=
paddle
.
trainer
.
SGD
(
cost
=
cost
,
parameters
=
parameters
,
update_equation
=
optimizer
)
return
avg_loss
```
### 读取数据且打印训练的中间信息
### 定义运算场所
我们可以定义运算是发生在CPU还是GPU
PaddlePaddle提供一个
[
reader机制
](
https://github.com/PaddlePaddle/Paddle/tree/develop/doc/design/reader
)
来读取数据。 Reader返回的数据可以包括多列,我们需要一个Python dict把列
序号映射到网络里的数据层。
```
python
use_cuda
=
False
place
=
fluid
.
CUDAPlace
(
0
)
if
use_cuda
else
fluid
.
CPUPlace
()
```
### 创建训练器
训练器会读入一个训练程序和一些必要的其他参数:
```
python
feeding
=
{
'x'
:
0
,
'y'
:
1
}
trainer
=
fluid
.
Trainer
(
train_func
=
train_program
,
place
=
place
,
optimizer_func
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
0.001
))
```
此外,我们还可以提供一个 event handler,来打印训练的进度:
### 开始提供数据
PaddlePaddle提供了读取数据者发生器机制来读取训练数据。读取数据者会一次提供多列数据,因此我们需要一个Python的list来定义读取顺序。
```
python
# event_handler to print training and testing info
def
event_handler
(
event
):
if
isinstance
(
event
,
paddle
.
event
.
EndIteration
):
if
event
.
batch_id
%
100
==
0
:
print
"Pass %d, Batch %d, Cost %f"
%
(
event
.
pass_id
,
event
.
batch_id
,
event
.
cost
)
if
isinstance
(
event
,
paddle
.
event
.
EndPass
):
result
=
trainer
.
test
(
reader
=
paddle
.
batch
(
uci_housing
.
test
(),
batch_size
=
2
),
feeding
=
feeding
)
print
"Test %d, Cost %f"
%
(
event
.
pass_id
,
result
.
cost
)
feed_order
=
[
'x'
,
'y'
]
```
除此之外,可以定义一个事件相应器来处理类似
`打印训练进程`
的事件:
```
python
#
event_handler to print training and testing info
from
paddle.v2.plot
import
Ploter
#
Specify the directory path to save the parameters
params_dirname
=
"fit_a_line.inference.model"
# Plot data
from
paddle.v2.plot
import
Ploter
train_title
=
"Train cost"
test_title
=
"Test cost"
cost_ploter
=
Ploter
(
train_title
,
test_title
)
plot_cost
=
Ploter
(
train_title
,
test_title
)
step
=
0
# event_handler to print training and testing info
def
event_handler_plot
(
event
):
global
step
if
isinstance
(
event
,
paddle
.
event
.
EndIteration
):
if
step
%
10
==
0
:
# every 10 batches, record a train cost
cost_ploter
.
append
(
train_title
,
step
,
event
.
cost
)
if
isinstance
(
event
,
fluid
.
EndStepEvent
):
if
event
.
step
%
10
==
0
:
# every 10 batches, record a test cost
test_metrics
=
trainer
.
test
(
reader
=
test_reader
,
feed_order
=
feed_order
)
if
step
%
100
==
0
:
# every 100 batches, record a test cost
result
=
trainer
.
test
(
reader
=
paddle
.
batch
(
uci_housing
.
test
(),
batch_size
=
2
),
feeding
=
feeding
)
cost_ploter
.
append
(
test_title
,
step
,
result
.
cost
)
plot_cost
.
append
(
test_title
,
step
,
test_metrics
[
0
])
plot_cost
.
plot
()
if
step
%
100
==
0
:
# every 100 batches, update cost plot
cost_ploter
.
plot
()
if
test_metrics
[
0
]
<
10.0
:
# If the accuracy is good enough, we can stop the training.
print
(
'loss is less than 10.0, stop'
)
trainer
.
stop
()
step
+=
1
# We can save the trained parameters for the inferences later
if
params_dirname
is
not
None
:
trainer
.
save_params
(
params_dirname
)
if
isinstance
(
event
,
paddle
.
event
.
EndPass
):
if
event
.
pass_id
%
10
==
0
:
with
open
(
'params_pass_%d.tar'
%
event
.
pass_id
,
'w'
)
as
f
:
trainer
.
save_parameter_to_tar
(
f
)
step
+=
1
```
### 开始训练
我们现在可以通过调用
`trainer.train()`
来开始训练
```
python
%
matplotlib
inline
# The training could take up to a few minutes.
trainer
.
train
(
reader
=
paddle
.
batch
(
paddle
.
reader
.
shuffle
(
uci_housing
.
train
(),
buf_size
=
500
),
batch_size
=
2
),
feeding
=
feeding
,
reader
=
train_reader
,
num_epochs
=
100
,
event_handler
=
event_handler_plot
,
num_passes
=
30
)
feed_order
=
feed_order
)
```
![
png
](
./image/train_and_test.png
)
### 应用模型
## 预测
提供一个
`inference_program`
和一个
`params_dirname`
来初始化预测器。
`params_dirname`
用来存储我们的参数。
### 设定预测程序
类似于
`trainer.train`
,预测器需要一个预测程序来做预测。我们可以稍加修改我们的训练程序来把预测值包含进来。
#### 1. 生成测试数据
```
python
test_data_creator
=
paddle
.
dataset
.
uci_housing
.
test
()
test_data
=
[]
test_label
=
[]
for
item
in
test_data_creator
():
test_data
.
append
((
item
[
0
],))
test_label
.
append
(
item
[
1
])
if
len
(
test_data
)
==
5
:
break
def
inference_program
():
x
=
fluid
.
layers
.
data
(
name
=
'x'
,
shape
=
[
13
],
dtype
=
'float32'
)
y_predict
=
fluid
.
layers
.
fc
(
input
=
x
,
size
=
1
,
act
=
None
)
return
y_predict
```
#### 2. 推测 inference
### 预测
预测器会从
`params_dirname`
中读取已经训练好的模型,来对从未遇见过的数据进行预测。
```
python
# load parameters from tar file.
# users can remove the comments and change the model name
# with open('params_pass_20.tar', 'r') as f:
# parameters = paddle.parameters.Parameters.from_tar(f)
inferencer
=
fluid
.
Inferencer
(
infer_func
=
inference_program
,
param_path
=
params_dirname
,
place
=
place
)
probs
=
paddle
.
infer
(
output_layer
=
y_predict
,
parameters
=
parameters
,
input
=
test_data
)
batch_size
=
10
tensor_x
=
numpy
.
random
.
uniform
(
0
,
10
,
[
batch_size
,
13
]).
astype
(
"float32"
)
for
i
in
xrange
(
len
(
probs
)):
print
"label="
+
str
(
test_label
[
i
][
0
])
+
", predict="
+
str
(
probs
[
i
]
[
0
])
results
=
inferencer
.
infer
({
'x'
:
tensor_x
})
print
(
"infer results: "
,
results
[
0
])
```
## 总结
...
...
01.fit_a_line/README.md
浏览文件 @
0fb04454
...
...
@@ -165,7 +165,7 @@ The trainer will take the `train_program` as input.
trainer
=
fluid
.
Trainer
(
train_func
=
train_program
,
place
=
place
,
optimizer
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
0.001
))
optimizer
_func
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
0.001
))
```
### Feeding Data
...
...
01.fit_a_line/train.py
浏览文件 @
0fb04454
...
...
@@ -47,7 +47,7 @@ place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
trainer
=
fluid
.
Trainer
(
train_func
=
train_program
,
place
=
place
,
optimizer
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
0.001
))
optimizer
_func
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
0.001
))
feed_order
=
[
'x'
,
'y'
]
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录