README.md 34.6 KB
Newer Older
C
choijulie 已提交
1
# Recognize Digits
Y
Yi Wang 已提交
2

H
Hao Wang 已提交
3
The source code of this tutorial is in [book/recognize_digits](https://github.com/PaddlePaddle/book/tree/develop/02.recognize_digits). For new users, please refer to [Running This Book](https://github.com/PaddlePaddle/book/blob/develop/README.md#running-the-book) .
Y
Yi Wang 已提交
4

H
Hao Wang 已提交
5 6
## Background
When we learning programming, the first program we write is generally to implement printing “Hello World”. But the tutorial of machine learning or deep learning for the beginner is usually handwriting recognition on the [MNIST](http://yann.lecun.com/exdb/mnist/) database. Because handwriting recognition is a typical classification problem, relatively simple and the MNIST dataset is complete. MNIST dataset as a simple computer vision dataset contains a series of pictures and corresponding labels of handwriting digits. The picture is a 28x28 pixel matrix, and the label corresponds to 10 numbers from 0 to 9. Each picture has been normalized in size and centered in the position.
Y
Yi Wang 已提交
7 8

<p align="center">
H
Hao Wang 已提交
9 10
<img src="https://github.com/PaddlePaddle/book/blob/develop/02.recognize_digits/image/mnist_example_image.png?raw=true" width="400"><br/>
Figure 1. Example of a MNIST picture
Y
Yi Wang 已提交
11 12
</p>

H
Hao Wang 已提交
13
MNIST dataset is created from [NIST](https://www.nist.gov/srd/nist-special-database-19) Special Database 3(SD-3) and Special Database 1(SD-1). Because SD-3 is labeled by stuff of US Census Bureau and SD-1 is labeled by US high school students, so SD-3 is clearer and easier to be recognized than SD-1. Yann LeCun et al. pick half of SD-1 and half of SD-3 as train dataset (60000 data) and test dataset (10000 data).250 annotators labeled the training set, thus guaranteed that there wasn't a complete overlap of annotators of training set and test set.
Y
Yi Wang 已提交
14

H
Hao Wang 已提交
15
MNIST attracts scholars to train model based on the dataset. In 1998, LeCun conducted experiments respectively using Single layer classifier, Multilayer Perceptron and Multilayer convolutional neural network LeNet, constantly decreasing the error on test dataset ( from 12% to 0.7%)\[[1](#References)\]。 In the process of research, LeCun, the pioneer in the field of deep learning, came up with Convolutional Neural Network, largely improving the performance of handwriting recognition. After that, researchers take a large number of experiments based on K-Nearest Neighbors algorithm\[[2](#References)\], SVM\[[3](#References)\], Neural Network\[[4-7](#References)\] and Boosting method\[[8](#References)\] and so on, with multiple pre-processing methods(like distortion removal, noise removal, and blurring) to upgrade accuracy of recognition.
Y
Yi Wang 已提交
16

H
Hao Wang 已提交
17
Convolutional Neural Network plays an important role in the field of deep learning now. From simple LeNet proposed by Yann LeCun in early days to model VGGNet, GoogleNet, ResNet and so on in the ImageNet competition (please refer to [Image Classification](https://github.com/PaddlePaddle/book/tree/develop/03.image_classification) tutorial ), we have gain a serious of great achievements with convolutional neural network in the field of image classification.
Y
Yi Wang 已提交
18 19 20



H
Hao Wang 已提交
21 22 23 24 25 26 27 28 29 30 31
In this tutorial, starting from simple Softmax regression model, we help you learn about handwriting recognition and introduce you how to upgrade model and how to use MLP and CNN to optimize recognition result.


## Exploration of Models

To train a classifier based on MNIST dataset, before the introduction of three basic image classification networks used in this tutorial, we first give some definitions:

- $X$ is the input: the MNIST image is a two-dimensional image of $28\times28$. For the calculation, we transform it into a $784$ dimensional vector, ie $X=\left ( x_0, x_1, \dots, x_{783} \right )$.

- $Y$ is the output: the output of the classifier is number (0-9), ie $Y=\left ( y_0, y_1, \dots, y_9 \right )$, and each dimension $y_i$ represents the probability of image classification as $i$th number.

X
xiaoting 已提交
32
- $Label$ is the actual label of the picture: $Label=\left ( l_0, l_1, \dots, l_9 \right ) $ is also 10 dimensions, but only one dimension represents 1, and the rest is 0. For example, if the number on an image is 2, its label is $(0,0,1,0, \dots, 0)$
Y
Yi Wang 已提交
33

C
choijulie 已提交
34
### Softmax Regression
Y
Yi Wang 已提交
35

H
Hao Wang 已提交
36
The simplest Softmax regression model is to get features with input layer passing through a fully connected layer and then compute and ouput probabilities of multiple classifications directly via Softmax function \[[9](#references)\].
Y
Yi Wang 已提交
37

H
Hao Wang 已提交
38
The data of the input layer $X$ is passed to the output layer. The input $X$ is multiplied by weights $W$ and then added to the bias $b$ to generate activations:
Y
Yi Wang 已提交
39

L
Luo Tao 已提交
40
$$ y_i = \text{softmax}(\sum_j W_{i,j}x_j + b_i) $$
Y
Yi Wang 已提交
41

C
choijulie 已提交
42
where $ \text{softmax}(x_i) = \frac{e^{x_i}}{\sum_j e^{x_j}} $
Y
Yi Wang 已提交
43

H
Hao Wang 已提交
44
Figure 2 is a network of softmax regression, in which weights are represented by blue lines, bias are represented by red lines, and +1 indicates that the bias is $1$.
Y
Yi Wang 已提交
45

H
Hao Wang 已提交
46 47 48 49
<p align="center">
<img src="https://github.com/PaddlePaddle/book/blob/develop/02.recognize_digits/image/softmax_regression.png?raw=true" width=400><br/>
Figure 2. Softmax regression network structure <br/>
</p>
Y
Yi Wang 已提交
50

H
Hao Wang 已提交
51 52 53 54 55
For an $N$-class classification problem with $N$ output nodes, Softmax normalizes the resulting $N$ dimensional vector so that each of its entries falls in the range $[0,1]\in {R}$, representing the probability that the sample belongs to a certain category. Here $y_i$ denotes the predicted probability that an image is of number $i$.

In the classification problem, we usually use cross-entropy loss, the formula is as follows:

$$  L_{cross-entropy}(label, y) = -\sum_i label_ilog(y_i) $$
Y
Yi Wang 已提交
56 57 58



C
choijulie 已提交
59
### Multilayer Perceptron
Y
Yi Wang 已提交
60

H
Hao Wang 已提交
61
The Softmax regression model uses the simplest two-layer neural network, which contains only the input layer and the output layer, so its performance is limited. In order to achieve better recognition, we consider adding several hidden layers \[[10](#references)\] between the input and output layer.
Y
Yi Wang 已提交
62

H
Hao Wang 已提交
63 64 65
1.In the first hidden layer, you can get $ H_1 = \phi(W_1X + b_1) $, where $\phi$ represents the activation function. And common functions are [sigmoid, tanh or ReLU](#common activation functions).
2.In the second hidden layer, you can get $ H_2 = \phi(W_2H_1 + b_2) $.
3.Finally, in the output layer, you can get $Y=\text{softmax}(W_3H_2 + b_3)$, that is the final classification result vector.
66

Y
Yi Wang 已提交
67

H
Hao Wang 已提交
68
Figure 3 is a network structure of a multi-layer perceptron, in which weights are represented by blue lines, bias are represented by red lines, and +1 indicates that the bias is $1$.
D
dangqingqing 已提交
69

H
Hao Wang 已提交
70 71 72
<p align="center">
<img src="https://github.com/PaddlePaddle/book/blob/develop/02.recognize_digits/image/mlp.png?raw=true" width=500><br/>
Figure 3. Multilayer perceptron network structure <br/>
D
dangqingqing 已提交
73 74
</p>

C
choijulie 已提交
75
### Convolutional Neural Network
D
dangqingqing 已提交
76

H
Hao Wang 已提交
77
In the multi-layer perceptron model, an image is expanded into a one-dimensional vector and input into the network, ignoring its position and structure information. And the convolutional neural network can better utilize the structure information of the image. [LeNet-5](http://yann.lecun.com/exdb/lenet/) is a relatively simple convolutional neural network. Figure 4 shows the structure: the input two-dimensional image, first through the two convolutional layers to the pooling layer, then through the fully connected layer, and finally using the softmax as the output layer. Below we mainly introduce the convolutional layer and the pooling layer.
D
dangqingqing 已提交
78 79

<p align="center">
H
Hao Wang 已提交
80 81
<img src="https://github.com/PaddlePaddle/book/blob/develop/02.recognize_digits/image/cnn.png?raw=true" width="600"><br/>
Figure 4. LeNet-5 convolutional neural network structure<br/>
D
dangqingqing 已提交
82 83
</p>

H
Hao Wang 已提交
84
#### Convolutional Layer
D
dangqingqing 已提交
85

H
Hao Wang 已提交
86
Convolutional Layer is the core of convolutional neural network. The convolution we mentioned in image recognition is a two-dimensional convolution, that is, a discrete two-dimensional filter (also called a convolutional kernel) and a two-dimensional image for convoluting. In short, the two-dimensional filter slides to all positions on two-dimensional images and dot product is taken for this pixel and its domain pixel at each position. Convolution operations are widely used in the field of image processing. Different convolutional kernels can extract different features, such as edges, lines, and angles. In deep convolutional neural networks, low-level to complex image features can be extracted by convolution operation.
D
dangqingqing 已提交
87

C
choijulie 已提交
88
<p align="center">
H
Hao Wang 已提交
89 90
<img src="https://github.com/PaddlePaddle/book/blob/develop/02.recognize_digits/image/conv_layer.png?raw=true" width='750'><br/>
Figure 5. Convolutional Layer Picture <br/>
C
choijulie 已提交
91
</p>
D
dangqingqing 已提交
92

H
Hao Wang 已提交
93 94 95 96 97 98 99 100
Figure 5 shows an example of the process of computing convolution with input image in size of $H=5, W=5, D=3$, ie $5 \times 5$ size of 3 channel (RGB, also known as depth) color image.

This example contains two (denoted by $K$) groups of convolutional kernels, i.e. $Filter W_0$ and $Filter W_1$ in the figure. In convolution calculation, different convolutional kernels are usually used for different input channels. In the example, each set of convolutional kernels contains ($D=3$) $3\times 3$ (indicated by $F \times F$) convolutional kernel. In addition, the stride of convolutional kernel in horizontal and vertical direction of image is 2 (indicated by $S$); Pad 1 (represented by $P$) 0 in the four directions of input image, that is, the input layer raw data in the figure is the blue part, and the gray part is expanded with 0 in the size of 1. The convolution operation yields a feature map of the size of $3 \times 3 \times 2$ (represented by $H_{o} \times W_{o} \times K$), which is a 2-channel feature map in size of $3 \times 3$, where $H_o$ is calculated as: $H_o = (H - F + 2 \times P)/S + 1$, so is $W_o$. And each pixel in the output feature map is the summation of the inner product of each set of filters and each feature of the input image, plus the bias $b_o$, the bias is usually shared by each output feature map. The last $-2$ in the output feature map $o[:,:,0]$ is calculated as shown in the lower right corner of Figure 5.

The convolutional kernel is a learnable parameter in the convolution operation. As shown in the example above, the parameter of each layer of convolution is $D \times F \times F \times K$. In the multi-layer perceptron model, neurons are usually fully connected therefore with a large number of parameters. There are fewer parameters in the convolutional layer, which is also determined by main features of the convolutional layer, namely local connections and shared weights.

- Local connection: Each neuron is connected to only one region of the input neuron, which is called Receptive Field. In the image convolution operation, that is, the neurons are locally connected in the spatial dimension (the plane in which the above examples H and W are located), but are fully connected in depth. For the two-dimensional image itself, the local pixels are strongly related. This local connection ensures that the learned filter makes the strongest response to local input features. The idea of local connection is also inspired by the structure of visual system in biology. The neurons in the visual cortex receive information locally.

X
xiaoting 已提交
101
- Weight sharing: The filters used to calculate neurons in the same deep slice are shared. For example, in Figure 5, the filter for each neuron calculated by $o[:,:,0]$ is the same, both are $W_0$, which can greatly reduce the parameters. The sharing weight is meaningful to a certain extent, for example, the bottom edge feature of the image is independent of the specific location of the feature in the graph. However, it is unintentional in some cases. For example, the input picture is a face, eyes and hair are in different positions. And to learn different features in different positions, please (refer to [Stanford University Open Class](http://cs231n.Github.io/convolutional-networks/)). Note that the weights are only shared for the neurons of the same depth slice. In the convolutional layer, multiple sets of convolutional kernels are usually used to extract different features, that is, the weights of neurons with different depth slices are not shared by the features with different depth slices. In addition, bias are shared by all neurons with the same depth.
D
dangqingqing 已提交
102

H
Hao Wang 已提交
103 104 105 106 107
By introducing the calculation process of convolution and its features, convolution could be seen as a linear operation with shift-invariant, which is the same operation performed at each position of the image. The local connection and weight sharing of the convolutional layer greatly reduce the parameters that need to be learned, which helps with training larger convolutional neural networks.

For more information about convolution, please refer to [Reference Reading](http://ufldl.stanford.edu/wiki/index.php/Feature_extraction_using_convolution#Convolutions)

### Pooling Layer
Y
Yi Wang 已提交
108 109

<p align="center">
H
Hao Wang 已提交
110 111
<img src="https://github.com/PaddlePaddle/book/blob/develop/02.recognize_digits/image/max_pooling.png?raw=true" width="400px"><br/>
Figure 6. Picture in pooling layer</br>
Y
Yi Wang 已提交
112 113
</p>

H
Hao Wang 已提交
114
Pooling is a form of nonlinear downsampling. The main functionality of this layer is to reduce computation by reducing the network parameters and to control the overfitting to some extent. Normally a pooling layer is added after the convolutional layer. Pooling includes maximum pooling, average pooling and so on. The largest pooling is to divide the input layer into different areas by non-overlapping rectangular boxes, and the maximum value of each rectangular box is taken as the output layer, as shown in Figure. 6.
C
choijulie 已提交
115

H
Hao Wang 已提交
116
For details about convolutional neural network, please refer to the tutorial of [Standford Online Course]( http://cs231n.github.io/convolutional-networks/ ), [Ufldl](http://ufldl.stanford.edu/wiki/index.php/Pooling) and [Image Classification]( https://github.com/PaddlePaddle/book/tree/develop/03.image_classification ).
C
choijulie 已提交
117

H
Hao Wang 已提交
118 119
<a name="common activation functions"></a>
### Common activation functions
Y
Yi Wang 已提交
120

C
choijulie 已提交
121
- Sigmoid activation function: $ f(x) = sigmoid(x) = \frac{1}{1+e^{-x}} $
Y
Yi Wang 已提交
122

C
choijulie 已提交
123
- Tanh activation function: $ f(x) = tanh(x) = \frac{e^x-e^{-x}}{e^x+e^{-x}} $
Y
Yi Wang 已提交
124

H
Hao Wang 已提交
125
In fact, the tanh function is only a sigmoid function with change of scale. The value of the sigmoid function is doubled and then shifted down by 1 unit: tanh(x) = 2sigmoid(2x) - 1 .
Y
Yi Wang 已提交
126

C
choijulie 已提交
127
- ReLU activation function: $ f(x) = max(0, x) $
Y
Yi Wang 已提交
128

H
Hao Wang 已提交
129
For details, please refer to [activation function in Wikipedia](https://en.wikipedia.org/wiki/Activation_function).
Y
Yi Wang 已提交
130

H
Hao Wang 已提交
131
## Dataset Preparation
Y
Yi Wang 已提交
132

H
Hao Wang 已提交
133
PaddlePaddle provides a module `paddle.dataset.mnist` that automatically loads [MNIST] (http://yann.lecun.com/exdb/mnist/) data in the API. The loaded data is located under `/home/username/.cache/paddle/dataset/mnist`:
Y
Yi Wang 已提交
134 135


H
Hao Wang 已提交
136 137 138 139 140 141
|    filename          |       note              |
|----------------------|-------------------------|
|train-images-idx3-ubyte|  train data picture, 60,000 data |
|train-labels-idx1-ubyte|  train data label, 60,000 data    |
|t10k-images-idx3-ubyte |  test data picture, 10,000 data  |
|t10k-labels-idx1-ubyte |  test data label, 10,000 data     |
Y
Yi Wang 已提交
142

W
Wang,Jeff 已提交
143 144
## Fluid API Overview

H
Hao Wang 已提交
145 146 147 148 149 150 151 152 153 154
The demo will use the latest [Fluid API](http://paddlepaddle.org/documentation/docs/en/1.2/api_cn/index_cn.html). Fluid API is the latest PaddlePaddle API. It simplifies model configuration without sacrificing performance.
We recommend using the Fluid API, which is easy to learn and use to help you complete your machine learning tasks quickly.

Here is an overview of several important concepts in the Fluid API:

1. `inference_program`: specifies how to get the inference function from the data input.
This is where the network flow is defined.

2. `train_program`: specifies how to get the `loss` function from `inference_program` and `tag value`.
This is where the loss calculation is specified.
W
Wang,Jeff 已提交
155

H
Hao Wang 已提交
156
3. `optimizer_func`: Specifies the function of the optimizer configuration. The optimizer is responsible for reducing losses and driving training. Paddle supports a number of different optimizers.
W
Wang,Jeff 已提交
157

H
Hao Wang 已提交
158
In the code examples below, we'll take a closer look at them.
W
Wang,Jeff 已提交
159

H
Hao Wang 已提交
160
## Configuration Instructions
Y
Yi Wang 已提交
161

H
Hao Wang 已提交
162
Load the Fluid API package for PaddlePaddle.
Y
Yi Wang 已提交
163 164

```python
165
from __future__ import print_function #load print of python3 into current version
H
Hao Wang 已提交
166 167 168 169 170
import os
from PIL import Image # load module of image processing
import matplotlib.pyplot as plt
import numpy
import paddle # load paddle module
171
import paddle.fluid as fluid
Y
Yi Wang 已提交
172 173
```

W
Wang,Jeff 已提交
174 175
### Program Functions Configuration

H
Hao Wang 已提交
176
We need to configure `inference_program` function. We want to use this program to show three different classifiers, each of which is defined as a Python function.
X
xiaoting 已提交
177
We need to input the image data into the classifier. Paddle provides a special layer `fluid.data` for reading data.
H
Hao Wang 已提交
178
Let's create a data layer to read the image and connect it to the network of classification.
C
choijulie 已提交
179

H
Hao Wang 已提交
180
-Softmax regression: The results of the classification can be obtained only through a simple layer of simple fully connected layer with softmax as the activation function.
Y
Yi Wang 已提交
181 182

```python
183
def softmax_regression():
H
Hao Wang 已提交
184 185 186 187 188 189 190
    """
    Define softmax classifier:
    A fully connected layer with activation function  softmax
    Return:
    predict_image -- result of classification
    """
    # input original image data in size of 28*28*1
C
ceci3 已提交
191
    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
H
Hao Wang 已提交
192
    # With softmax as the fully connected layer of the activation function, the size of the output layer must be 10
W
Wang,Jeff 已提交
193
    predict = fluid.layers.fc(
H
Hao Wang 已提交
194
    input=img, size=10, act='softmax')
Y
Yi Wang 已提交
195 196
    return predict
```
C
choijulie 已提交
197

H
Hao Wang 已提交
198
-Multilayer Perceptron: The following code implements a multilayer perceptron with two hidden layers (that is, fully connected layers). The activation functions of the two hidden layers are all ReLU, and the activation function of the output layer is Softmax.
Y
Yi Wang 已提交
199 200

```python
201
def multilayer_perceptron():
H
Hao Wang 已提交
202 203 204 205 206 207 208 209 210
    """
    Define multilayer perceptron classifier:
    Multilayer perceptron with two hidden layers (fully connected layers)
    The activation function of the first two hidden layers uses ReLU, and the activation function of the output layer uses Softmax.

    Return:
    predict_image -- result of classification
    """
    # input raw image data in size of 28*28*1
C
ceci3 已提交
211
    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
H
Hao Wang 已提交
212
    # the first fully connected layer, whose activation function is ReLU
213
    hidden = fluid.layers.fc(input=img, size=200, act='relu')
H
Hao Wang 已提交
214
    # the second fully connected layer, whose activation function is ReLU
215
    hidden = fluid.layers.fc(input=hidden, size=200, act='relu')
H
Hao Wang 已提交
216
    # With softmax as the fully connected output layer of the activation function, the size of the output layer must be 10
217 218
    prediction = fluid.layers.fc(input=hidden, size=10, act='softmax')
    return prediction
Y
Yi Wang 已提交
219
```
C
choijulie 已提交
220

H
Hao Wang 已提交
221
-Convolutional neural network LeNet-5: The input two-dimensional image first passes through two convolutional layers to the pooling layer, then passes through the fully connected layer, and finally fully connection layer with softmax as activation function is used as output layer.
Y
Yi Wang 已提交
222 223

```python
224
def convolutional_neural_network():
H
Hao Wang 已提交
225 226 227 228 229 230 231 232
    """
    Define convolutional neural network classifier:
    The input 2D image passes through two convolution-pooling layers, using the fully connected layer with softmax as the output layer

    Return:
    predict -- result of classification
    """
    # input raw image data in size of 28*28*1
C
ceci3 已提交
233
    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
H
Hao Wang 已提交
234 235
    # the first convolution-pooling layer
    # Use 20 5*5 filters, the pooling size is 2, the pooling step is 2, and the activation function is Relu.
236
    conv_pool_1 = fluid.nets.simple_img_conv_pool(
Y
Yi Wang 已提交
237 238 239 240 241
        input=img,
        filter_size=5,
        num_filters=20,
        pool_size=2,
        pool_stride=2,
242 243
        act="relu")
    conv_pool_1 = fluid.layers.batch_norm(conv_pool_1)
H
Hao Wang 已提交
244 245
    # the second convolution-pooling layer
    # Use 20 5*5 filters, the pooling size is 2, the pooling step is 2, and the activation function is Relu.
246
    conv_pool_2 = fluid.nets.simple_img_conv_pool(
Y
Yi Wang 已提交
247 248 249 250 251
        input=conv_pool_1,
        filter_size=5,
        num_filters=50,
        pool_size=2,
        pool_stride=2,
252
        act="relu")
H
Hao Wang 已提交
253
    # With softmax as the fully connected output layer of the activation function, the size of the output layer must be 10
254 255
    prediction = fluid.layers.fc(input=conv_pool_2, size=10, act='softmax')
    return prediction
Y
Yi Wang 已提交
256 257
```

W
Wang,Jeff 已提交
258
#### Train Program Configuration
H
Hao Wang 已提交
259 260
Then we need to set train program `train_program` It firstly infers from classifier.
During the training, it will compute `avg_cost`.
261

H
Hao Wang 已提交
262
** Note:** train program should return an array. The first parameter returned must be `avg_cost`. The trainer uses it to compute gradient.
263

H
Hao Wang 已提交
264
Please write your code and then test results of different classifiers of  `softmax_regression`, `MLP` and `convolutional neural network`.
Y
Yi Wang 已提交
265 266

```python
267
def train_program():
H
Hao Wang 已提交
268 269 270 271 272 273 274 275 276 277
    """
    Configure train_program

    Return:
    predict -- result of classification
    avg_cost -- mean loss
    acc -- accuracy of classification

    """
    # label layer, called label, correspondent with label category of input picture
C
ceci3 已提交
278
    label = fluid.data(name='label', shape=[None, 1], dtype='int64')
279

H
Hao Wang 已提交
280 281 282
    # predict = softmax_regression() # cancel note and run Softmax regression
    # predict = multilayer_perceptron() # cancel note and run multiple perceptron
    predict = convolutional_neural_network() # cancel note and run LeNet5 convolutional neural network
W
Wang,Jeff 已提交
283

H
Hao Wang 已提交
284
    # use class cross-entropy function to compute loss function between predict and label
285
    cost = fluid.layers.cross_entropy(input=predict, label=label)
H
Hao Wang 已提交
286
    # compute mean loss
287
    avg_cost = fluid.layers.mean(cost)
H
Hao Wang 已提交
288
    # compute accuracy of classification
289
    acc = fluid.layers.accuracy(input=predict, label=label)
H
Hao Wang 已提交
290
    return predict, [avg_cost, acc]
291

Y
Yi Wang 已提交
292 293
```

294 295
#### Optimizer Function Configuration

H
Hao Wang 已提交
296
`Adam optimizer``learning_rate` below are learning rate. Their size is associated with speed of network train convergence.
297 298 299 300 301 302

```python
def optimizer_program():
    return fluid.optimizer.Adam(learning_rate=0.001)
```

H
Hao Wang 已提交
303
### Data Feeders for dataset Configuration
Y
Yi Wang 已提交
304

H
Hao Wang 已提交
305
Next We start the training process. `Paddle.dataset.mnist.train()` and `paddle.dataset.mnist.test()` are respectively as train dataset and test dataset. These two functions respectively return a reader-- reader in PaddlePaddle is a Python function, which returns a Python yield generator when calling the reader.
Y
Yi Wang 已提交
306

H
Hao Wang 已提交
307
`Shuffle` below is a reader decorator, which receives a reader A and returns another reader B. Reader B read `buffer_size` train data into a buffer and then the data is disordered randomly and is output one by one.
Y
Yi Wang 已提交
308

H
Hao Wang 已提交
309
`Batch` is a special decorator. Its input is a reader and output is a batched reader. In PaddlePaddle, a reader yield a piece of data every time while batched reader yield a minibatch every time.
Y
Yi Wang 已提交
310

Q
qiaolongfei 已提交
311
```python
H
Hao Wang 已提交
312 313 314 315
# there are 64 data in a minibatch
BATCH_SIZE = 64

# read 500 data in train dataset, randomly disorder them and then transfer it into batched reader which yield 64 data each time.
316 317 318
train_reader = paddle.batch(
        paddle.reader.shuffle(
            paddle.dataset.mnist.train(), buf_size=500),
H
Hao Wang 已提交
319 320
        batch_size=BATCH_SIZE)
# read data in test dataset and yield 64 data every time
321
test_reader = paddle.batch(
H
Hao Wang 已提交
322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
            paddle.dataset.mnist.test(), batch_size=BATCH_SIZE)
```

### create training process

Now we need to create a training process. We will use `train_program`, `place` and `optimizer` defined before, conclude test loss in the period of training iteration and training verification and save parameters of model for prediction.


#### Event Handler Configuration

We can call a handler function to supervise training process during training.
We display two `event_handler` programs here. Please freely update Jupyter Notebook and find the changes.

`Event_handler` is used to output training result during the train

```python
def event_handler(pass_id, batch_id, cost):
    # print the intermediate results of training, like
    # training iterations, number of batch, and loss function
    print("Pass %d, Batch %d, Cost %f" % (pass_id,batch_id, cost))
```

```python
from paddle.utils.plot import Ploter

train_prompt = "Train cost"
test_prompt = "Test cost"
cost_ploter = Ploter(train_prompt, test_prompt)

# visualize training process
def event_handler_plot(ploter_title, step, cost):
    cost_ploter.append(ploter_title, step, cost)
    cost_ploter.plot()
Q
qiaolongfei 已提交
355 356
```

H
Hao Wang 已提交
357 358 359 360
`event_handler_plot` can be visualized as follows:

![png](./image/train_and_test.png)

W
Wang,Jeff 已提交
361

H
Hao Wang 已提交
362 363 364 365 366 367 368 369
### Start training

Aftering adding `event_handler` and `data reader` we configured, we can start to train the model.
Set parameters for operation to configure data description.
`Feed_order` is used to map data directory to `train_program`
Create a `train_test` reflecting the loss during our training.

Define network structure:
C
choijulie 已提交
370

Y
Yi Wang 已提交
371
```python
H
Hao Wang 已提交
372 373
# the model is run on single CPU
use_cuda = False # If you want to use GPU, please set it True
374
place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
W
Wang,Jeff 已提交
375

H
Hao Wang 已提交
376 377
# call train_program to get prediction value and loss value,
prediction, [avg_loss, acc] = train_program()
Y
Yi Wang 已提交
378

H
Hao Wang 已提交
379 380
# input original image data in size of 28*28*1
# label layer, called label, correspondent with label category of input picture.
X
xiaoting 已提交
381

H
Hao Wang 已提交
382
# It is informed that data in network consists of two parts. One is img value, the other is label value.
X
xiaoting 已提交
383
feeder = fluid.DataFeeder(feed_list=['img', 'label'], place=place)
W
Wang,Jeff 已提交
384

H
Hao Wang 已提交
385
# choose Adam optimizer
X
xiaoting 已提交
386
optimizer = optimizer_program()
H
Hao Wang 已提交
387 388
optimizer.minimize(avg_loss)
```
389

H
Hao Wang 已提交
390
Configure hyper parameter during the training:
W
Wang,Jeff 已提交
391 392

```python
H
Hao Wang 已提交
393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436

PASS_NUM = 5 #train 5 iterations
epochs = [epoch_id for epoch_id in range(PASS_NUM)]

# save parameters of model into save_dirname file
save_dirname = "recognize_digits.inference.model"
```


```python
def train_test(train_test_program,
                   train_test_feed, train_test_reader):

    # save classification accuracy into acc_set
    acc_set = []
    # save mean loss in avg_loss_set
    avg_loss_set = []
    # transfer each data which is the output of testing reader_yield into network to train
    for test_data in train_test_reader():
        acc_np, avg_loss_np = exe.run(
            program=train_test_program,
            feed=train_test_feed.feed(test_data),
            fetch_list=[acc, avg_loss])
        acc_set.append(float(acc_np))
        avg_loss_set.append(float(avg_loss_np))
    # get accuracy and loss value on the test data
    acc_val_mean = numpy.array(acc_set).mean()
    avg_loss_val_mean = numpy.array(avg_loss_set).mean()
    # return mean loss value and mean accuracy
    return avg_loss_val_mean, acc_val_mean
```

Create executor

```python
exe = fluid.Executor(place)
exe.run(fluid.default_startup_program())
```

Set up main_program and test_program:

```python
main_program = fluid.default_main_program()
test_program = fluid.default_main_program().clone(for_test=True)
437 438
```

H
Hao Wang 已提交
439 440
Start training:

Y
Yi Wang 已提交
441

442 443


Q
qiaolongfei 已提交
444
```python
H
Hao Wang 已提交
445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
lists = []
step = 0
for epoch_id in epochs:
    for step_id, data in enumerate(train_reader()):
        metrics = exe.run(main_program,
                          feed=feeder.feed(data),
                          fetch_list=[avg_loss, acc])
        if step % 100 == 0: # print a log for every 100 times of training
            print("Pass %d, Batch %d, Cost %f" % (step, epoch_id, metrics[0]))
            event_handler_plot(train_prompt, step, metrics[0])
        step += 1

    # test classification result of each epoch
    avg_loss_val, acc_val = train_test(train_test_program=test_program,
                                       train_test_reader=test_reader,
                                       train_test_feed=feeder)

    print("Test with Epoch %d, avg_cost: %s, acc: %s" %(epoch_id, avg_loss_val, acc_val))
    event_handler_plot(test_prompt, step, metrics[0])

    lists.append((epoch_id, avg_loss_val, acc_val))

    # save parameters of trained model for prediction
    if save_dirname is not None:
        fluid.io.save_inference_model(save_dirname,
                                      ["img"], [prediction], exe,
                                      model_filename=None,
                                      params_filename=None)

# Choose the best pass
best = sorted(lists, key=lambda list: float(list[1]))[0]
print('Best pass is %s, testing Avgcost is %s' % (best[0], best[1]))
print('The classification accuracy is %.2f%%' % (float(best[2]) * 100))
Y
Yi Wang 已提交
478 479
```

H
Hao Wang 已提交
480 481 482 483 484 485

The training process is completely automatic. The log printed in event_handler is like as follows.

Pass represents iterations of train. Batch represents times to train all data. cost represents loss value of current pass.

Compute the mean loss and accuracy of classification after an epoch.
Y
Yi Wang 已提交
486

487 488 489 490 491 492 493 494 495 496 497 498 499
```
Pass 0, Batch 0, Cost 0.125650
Pass 100, Batch 0, Cost 0.161387
Pass 200, Batch 0, Cost 0.040036
Pass 300, Batch 0, Cost 0.023391
Pass 400, Batch 0, Cost 0.005856
Pass 500, Batch 0, Cost 0.003315
Pass 600, Batch 0, Cost 0.009977
Pass 700, Batch 0, Cost 0.020959
Pass 800, Batch 0, Cost 0.105560
Pass 900, Batch 0, Cost 0.239809
Test with Epoch 0, avg_cost: 0.053097883707459624, acc: 0.9822850318471338
```
Y
Yi Wang 已提交
500

H
Hao Wang 已提交
501
Check prediction accuracy of the model after training. In the train with MNIST, generally classification accuracy of softmax regression model is about 92.34%, while that of multilayer perceptron is 97.66% and that of convolutional neural network is 99.20%.
Y
Yi Wang 已提交
502

C
choijulie 已提交
503

H
Hao Wang 已提交
504
## Deploy the Model
L
liaogang 已提交
505

H
Hao Wang 已提交
506
You can use trained model to classify handwriting pictures of digits. The program below shows how to use well-trained model to predict.
507

H
Hao Wang 已提交
508
### Generate input data to be inferred
509

H
Hao Wang 已提交
510
`infer_3.png` is an example picture of number 3. Transform it into a numpy to match feed data format
511 512


L
liaogang 已提交
513
```python
514
def load_image(file):
X
xiaoting 已提交
515
    # open the image file and covert to grayscale
516
    im = Image.open(file).convert('L')
X
xiaoting 已提交
517
    # adjust the input image to a 28*28 high quality image
518
    im = im.resize((28, 28), Image.ANTIALIAS)
X
xiaoting 已提交
519
    # convert img to numpy
H
Hao Wang 已提交
520
    im = numpy.array(im).reshape(1, 1, 28, 28).astype(numpy.float32)
X
xiaoting 已提交
521
    # normalize
522 523 524
    im = im / 255.0 * 2.0 - 1.0
    return im

525
cur_dir = os.getcwd()
H
Hao Wang 已提交
526
tensor_img = load_image(cur_dir + '/image/infer_3.png')
527
```
528

529
### Inference
530

H
Hao Wang 已提交
531 532 533
By configuring network and training parameters via `load_inference_model`, We can simply insert classifier defined before.


534

535
```python
H
Hao Wang 已提交
536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555
inference_scope = fluid.core.Scope()
with fluid.scope_guard(inference_scope):
    # use fluid.io.load_inference_model to get inference program desc,
    # feed_target_names is used to define variable name needed to be passed into network
    # fetch_targets define variable name to be fetched from network
    [inference_program, feed_target_names,
     fetch_targets] = fluid.io.load_inference_model(
     save_dirname, exe, None, None)

    # Make feed a dictionary {feed_target_name: feed_target_data}
    # The result will contain a data list corresponding to fetch_targets
    results = exe.run(inference_program,
                            feed={feed_target_names[0]: tensor_img},
                            fetch_list=fetch_targets)
    lab = numpy.argsort(results)

    # Print prediction result of  infer_3.png
    img=Image.open('image/infer_3.png')
    plt.imshow(img)
    print("Inference result of image/infer_3.png is: %d" % lab[0][0][-1])
L
liaogang 已提交
556 557
```

Y
Yi Wang 已提交
558

C
choijulie 已提交
559 560


H
Hao Wang 已提交
561 562 563 564
### Result

If successful, the inference result input is as follows:
`Inference result of image/infer_3.png is: 3` , which indicates that out network successfully recognize the picture!
C
choijulie 已提交
565

H
Hao Wang 已提交
566
## Summary
C
choijulie 已提交
567

H
Hao Wang 已提交
568
Softmax regression, multilayer perceptron and convolutional neural network are the most basic deep learning model, from which complex neural networks are all derivative, so these models are helpful for later learning. At the same time, we found that from simple softmax regression transform to slightly complex convolutional neural network, the accuracy of recognition on MNIST dataset largely increased, resulting from that convolution layer is featured with local connection and sharing weight. When study of new models later, hope you make a deep understand of the key upgrade of new model compared with original model. In addition, this tutorial also talks about the basic steps to build PaddlePadle model, from the code of dataprovider, build of network to training and prediction. Familiar with the work flow, you can use your own data, define your own network model and finish your training and prediction tasks.
Y
Yi Wang 已提交
569

H
Hao Wang 已提交
570
<a name="References"></a>
C
choijulie 已提交
571
## References
Y
Yi Wang 已提交
572 573

1. LeCun, Yann, Léon Bottou, Yoshua Bengio, and Patrick Haffner. ["Gradient-based learning applied to document recognition."](http://ieeexplore.ieee.org/abstract/document/726791/) Proceedings of the IEEE 86, no. 11 (1998): 2278-2324.
H
Hao Wang 已提交
574
2. Wejéus, Samuel. ["A Neural Network Approach to Arbitrary SymbolRecognition on Modern Smartphones."](http://www.diva-portal.org/smash/record.jsf?pid=diva2%3A753279&dswid=-434) (2014).
Y
Yi Wang 已提交
575 576 577 578 579 580 581
3. Decoste, Dennis, and Bernhard Schölkopf. ["Training invariant support vector machines."](http://link.springer.com/article/10.1023/A:1012454411458) Machine learning 46, no. 1-3 (2002): 161-190.
4. Simard, Patrice Y., David Steinkraus, and John C. Platt. ["Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis."](http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.160.8494&rep=rep1&type=pdf) In ICDAR, vol. 3, pp. 958-962. 2003.
5. Salakhutdinov, Ruslan, and Geoffrey E. Hinton. ["Learning a Nonlinear Embedding by Preserving Class Neighbourhood Structure."](http://www.jmlr.org/proceedings/papers/v2/salakhutdinov07a/salakhutdinov07a.pdf) In AISTATS, vol. 11. 2007.
6. Cireşan, Dan Claudiu, Ueli Meier, Luca Maria Gambardella, and Jürgen Schmidhuber. ["Deep, big, simple neural nets for handwritten digit recognition."](http://www.mitpressjournals.org/doi/abs/10.1162/NECO_a_00052) Neural computation 22, no. 12 (2010): 3207-3220.
7. Deng, Li, Michael L. Seltzer, Dong Yu, Alex Acero, Abdel-rahman Mohamed, and Geoffrey E. Hinton. ["Binary coding of speech spectrograms using a deep auto-encoder."](http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.185.1908&rep=rep1&type=pdf) In Interspeech, pp. 1692-1695. 2010.
8. Kégl, Balázs, and Róbert Busa-Fekete. ["Boosting products of base classifiers."](http://dl.acm.org/citation.cfm?id=1553439) In Proceedings of the 26th Annual International Conference on Machine Learning, pp. 497-504. ACM, 2009.
9. Rosenblatt, Frank. ["The perceptron: A probabilistic model for information storage and organization in the brain."](http://psycnet.apa.org/journals/rev/65/6/386/) Psychological review 65, no. 6 (1958): 386.
582
10. Bishop, Christopher M. ["Pattern recognition."](http://users.isr.ist.utl.pt/~wurmd/Livros/school/Bishop%20-%20Pattern%20Recognition%20And%20Machine%20Learning%20-%20Springer%20%202006.pdf) Machine Learning 128 (2006): 1-58.
Y
Yi Wang 已提交
583 584

<br/>
X
xiaoting 已提交
585
<a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/"><img alt="知识共享许可协议" style="border-width:0" src="https://paddlepaddleimage.cdn.bcebos.com/bookimage/camo.png" /></a><br /><span xmlns:dct="http://purl.org/dc/terms/" href="http://purl.org/dc/dcmitype/Text" property="dct:title" rel="dct:type">This tutorial</span> is contributed by <a xmlns:cc="http://creativecommons.org/ns#" href="http://book.paddlepaddle.org" property="cc:attributionName" rel="cc:attributionURL">PaddlePaddle</a>, and licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/">Creative Commons Attribution-ShareAlike 4.0 International License</a>.