MTCNN（九）更改python与c代码的PReLU为ReLU

秒速五厘米 2022-04-13 07:15 363阅读 0赞

**背景：**MTCNN的prelu需要训练，并且prelu需要有参数。prelu改为relu之后，相应的参数量会降低并且实现也更简单。

**目的：**更改MTCNN的prelu为ReLU,并在python端训练出相应的代码，c端结构改为与python端一致。

**参考相关：**

MTCNN（三）基于python代码的网络结构更改 [https://blog.csdn.net/weixin\_36474809/article/details/82856171][https_blog.csdn.net_weixin_36474809_article_details_82856171]

MTCNN（一）训练与运行  [https://blog.csdn.net/weixin\_36474809/article/details/82752199][https_blog.csdn.net_weixin_36474809_article_details_82752199]

**目录**

一、python端结构的更改与参数训练

1.1 更改的位置

1.2 实验结果

1.3 重新训练与测试

Pnet

Rnet

Onet

二、采用网络结构表

Pnet

Rnet

Onet

三、c代码端结构的更改

3.1 函数的定义

3.2 网络的更改

--------------------

# 一、python端结构的更改与参数训练 #

## 1.1 更改的位置 ##

更改前的prelu

@layer
        def prelu(self, inp, name):
            with tf.variable_scope(name):
                i = int(inp.get_shape()[-1])
                alpha = self.make_var('alpha', shape=(i,))
                return tf.nn.relu(inp) + tf.multiply(alpha, -tf.nn.relu(-inp))

加上relu

@layer
        def relu(self, inp, name):
            return tf.nn.relu(inp, name=name)

然后把pnet/rnet/onet中的prelu函数全部改为relu函数。

这里批评一下Notepad++之中的空格总是与python需要的空格不一致。每次编辑好之后还要用spyder再进行一次空格，免得报错。IndentationError: unexpected unindent

网络结构定义之中的更改：prelu改为了relu，（层的名字方便起见暂时没有更改）

class PNet(NetWork):
    
        def setup(self, task='data', reuse=False):
    
            with tf.variable_scope('pnet', reuse=reuse):
                (
                    self.feed(task) .conv(
                        3,
                        3,
                        10,
                        1,
                        1,
                        padding='VALID',
                        relu=False,
                        name='conv1') .relu(
                        name='ReLU1') .conv(
                        3,
                        3,
                        16,
                        2,
                        2,
                        padding='SAME',
                        relu=False,
                        name='pool1_conv1') .relu(
                        name='pool1_ReLU1') .conv(
                        3,
                        3,
                        32,
                        1,
                        1,
                        padding='VALID',
                        relu=False,
                        name='conv2') .relu(
                        name='ReLU2') .conv(
                        3,
                        3,
                        32,
                        1,
                        1,
                        task=task,
                        padding='VALID',
                        relu=False,
                        name='conv3',
                        wd=self.weight_decay_coeff) .relu(
                        name='ReLU3'))
    
            if self.mode == 'train':
                if task == 'cls':
                    (self.feed('ReLU3')
                         .conv(1, 1, 2, 1, 1, task=task, relu=False,
                               name='pnet/conv4-1', wd=self.weight_decay_coeff))
                elif task == 'bbx':
                    (self.feed('ReLU3')
                         .conv(1, 1, 4, 1, 1, task=task, relu=False,
                               name='pnet/conv4-2', wd=self.weight_decay_coeff))
                elif task == 'pts':
                    (self.feed('ReLU3')
                         .conv(1, 1, 10, 1, 1, task=task, relu=False,
                               name='pnet/conv4-3', wd=self.weight_decay_coeff))
                self.out_put.append(self.get_output())
            else:
                (self.feed('ReLU3')
                     .conv(1, 1, 2, 1, 1, relu=False, name='pnet/conv4-1')
                     .softmax(name='softmax'))
                self.out_put.append(self.get_output())
                (self.feed('ReLU3')
                     .conv(1, 1, 4, 1, 1, relu=False, name='pnet/conv4-2'))
                self.out_put.append(self.get_output())

## 1.2 实验结果 ##

初步结果：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl8zNjQ3NDgwOQ_size_16_color_FFFFFF_t_70][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl8zNjQ3NDgwOQ_size_16_color_FFFFFF_t_70 1][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl8zNjQ3NDgwOQ_size_16_color_FFFFFF_t_70 2][]

后续检验过的结果：mAP  57.85%

## 1.3 重新训练与测试 ##

### Pnet ###

把其中的prelu改为relu，然后训练。

all trainable variables:
 <tf.Variable 'pnet/conv1/weights:0' shape=(3, 3, 3, 10) dtype=float32_ref>
 <tf.Variable 'pnet/conv1/biases:0' shape=(10,) dtype=float32_ref>
 <tf.Variable 'pnet/pool1_conv1/weights:0' shape=(3, 3, 10, 16) dtype=float32_ref>
 <tf.Variable 'pnet/pool1_conv1/biases:0' shape=(16,) dtype=float32_ref>
 <tf.Variable 'pnet/conv2/weights:0' shape=(3, 3, 16, 32) dtype=float32_ref>
 <tf.Variable 'pnet/conv2/biases:0' shape=(32,) dtype=float32_ref>
 <tf.Variable 'pnet/conv3/weights:0' shape=(3, 3, 32, 32) dtype=float32_ref>
 <tf.Variable 'pnet/conv3/biases:0' shape=(32,) dtype=float32_ref>
 <tf.Variable 'pnet/conv4-1/weights:0' shape=(1, 1, 32, 2) dtype=float32_ref>
 <tf.Variable 'pnet/conv4-1/biases:0' shape=(2,) dtype=float32_ref>
 <tf.Variable 'pnet/conv4-2/weights:0' shape=(1, 1, 32, 4) dtype=float32_ref>
 <tf.Variable 'pnet/conv4-2/biases:0' shape=(4,) dtype=float32_ref>
 <tf.Variable 'pnet/conv4-3/weights:0' shape=(1, 1, 32, 10) dtype=float32_ref>
 <tf.Variable 'pnet/conv4-3/biases:0' shape=(10,) dtype=float32_ref>
 all local variable:
 input/input_producer/limit_epochs/epochs:0
 input_1/input_producer/limit_epochs/epochs:0

只有卷积核与bias需要训练，没有了prelu的斜率。

### Rnet ###

all trainable variables:
 <tf.Variable 'rnet/conv1/weights:0' shape=(3, 3, 3, 28) dtype=float32_ref>
 <tf.Variable 'rnet/conv1/biases:0' shape=(28,) dtype=float32_ref>
 <tf.Variable 'rnet/pool1_conv1/weights:0' shape=(3, 3, 28, 28) dtype=float32_ref>
 <tf.Variable 'rnet/pool1_conv1/biases:0' shape=(28,) dtype=float32_ref>
 <tf.Variable 'rnet/conv2/weights:0' shape=(3, 3, 28, 48) dtype=float32_ref>
 <tf.Variable 'rnet/conv2/biases:0' shape=(48,) dtype=float32_ref>
 <tf.Variable 'rnet/pool2_conv3/weights:0' shape=(3, 3, 48, 48) dtype=float32_ref>
 <tf.Variable 'rnet/pool2_conv3/biases:0' shape=(48,) dtype=float32_ref>
 <tf.Variable 'rnet/conv3/weights:0' shape=(3, 3, 48, 64) dtype=float32_ref>
 <tf.Variable 'rnet/conv3/biases:0' shape=(64,) dtype=float32_ref>
 <tf.Variable 'rnet/conv4/weights:0' shape=(576, 128) dtype=float32_ref>
 <tf.Variable 'rnet/conv4/biases:0' shape=(128,) dtype=float32_ref>
 <tf.Variable 'rnet/conv5-1/weights:0' shape=(128, 2) dtype=float32_ref>
 <tf.Variable 'rnet/conv5-1/biases:0' shape=(2,) dtype=float32_ref>
 <tf.Variable 'rnet/conv5-2/weights:0' shape=(128, 4) dtype=float32_ref>
 <tf.Variable 'rnet/conv5-2/biases:0' shape=(4,) dtype=float32_ref>
 <tf.Variable 'rnet/conv5-3/weights:0' shape=(128, 10) dtype=float32_ref>
 <tf.Variable 'rnet/conv5-3/biases:0' shape=(10,) dtype=float32_ref>
 all local variable:
 input/input_producer/limit_epochs/epochs:0

### Onet ###

all trainable variables:
 <tf.Variable 'onet/conv1/weights:0' shape=(3, 3, 3, 32) dtype=float32_ref>
 <tf.Variable 'onet/conv1/biases:0' shape=(32,) dtype=float32_ref>
 <tf.Variable 'onet/conv2/weights:0' shape=(3, 3, 32, 32) dtype=float32_ref>
 <tf.Variable 'onet/conv2/biases:0' shape=(32,) dtype=float32_ref>
 <tf.Variable 'onet/conv3/weights:0' shape=(3, 3, 32, 64) dtype=float32_ref>
 <tf.Variable 'onet/conv3/biases:0' shape=(64,) dtype=float32_ref>
 <tf.Variable 'onet/conv4_/weights:0' shape=(3, 3, 64, 64) dtype=float32_ref>
 <tf.Variable 'onet/conv4_/biases:0' shape=(64,) dtype=float32_ref>
 <tf.Variable 'onet/conv5_/weights:0' shape=(3, 3, 64, 128) dtype=float32_ref>
 <tf.Variable 'onet/conv5_/biases:0' shape=(128,) dtype=float32_ref>
 <tf.Variable 'onet/conv6_/weights:0' shape=(3, 3, 128, 128) dtype=float32_ref>
 <tf.Variable 'onet/conv6_/biases:0' shape=(128,) dtype=float32_ref>
 <tf.Variable 'onet/conv5/weights:0' shape=(1152, 256) dtype=float32_ref>
 <tf.Variable 'onet/conv5/biases:0' shape=(256,) dtype=float32_ref>
 <tf.Variable 'onet/conv6-1/weights:0' shape=(256, 2) dtype=float32_ref>
 <tf.Variable 'onet/conv6-1/biases:0' shape=(2,) dtype=float32_ref>
 <tf.Variable 'onet/conv6-2/weights:0' shape=(256, 4) dtype=float32_ref>
 <tf.Variable 'onet/conv6-2/biases:0' shape=(4,) dtype=float32_ref>
 <tf.Variable 'onet/conv6-3/weights:0' shape=(256, 10) dtype=float32_ref>
 <tf.Variable 'onet/conv6-3/biases:0' shape=(10,) dtype=float32_ref>
 all local variable:
 input/input_producer/limit_epochs/epochs:0

# 二、采用网络结构表 #

## Pnet ##

<table> 
 <tbody> 
 <tr> 
 <td> Feature size </td> 
 <td>name</td> 
 <td> Kernel size </td> 
 <td> Stride </td> 
 <td> Padding </td> 
 </tr> 
 <tr> 
 <td> 12*12*3 </td> 
 <td> conv1 ReLU1 </td> 
 <td> 3*3*10 </td> 
 <td> 1 </td> 
 <td>Valid</td> 
 </tr> 
 <tr> 
 <td> 10*10*10 </td> 
 <td>pool1_conv1 pool1_ReLU1 </td> 
 <td> 3*3*16 </td> 
 <td> 2 </td> 
 <td> Same </td> 
 </tr> 
 <tr> 
 <td> 5*5*16 </td> 
 <td> conv2 ReLU2 </td> 
 <td> 3*3*32 </td> 
 <td> 1 </td> 
 <td>Valid</td> 
 </tr> 
 <tr> 
 <td> 3*3*32 </td> 
 <td> conv3 ReLU3 </td> 
 <td> 3*3*32 </td> 
 <td> 1 </td> 
 <td> Valid </td> 
 </tr> 
 <tr> 
 <td> 1*1*32 </td> 
 <td>&nbsp;</td> 
 <td> &nbsp; </td> 
 <td> &nbsp; </td> 
 <td> &nbsp; </td> 
 </tr> 
 </tbody> 
</table>

## Rnet ##

<table> 
 <tbody> 
 <tr> 
 <td> Feature size </td> 
 <td>name</td> 
 <td> Kernel size </td> 
 <td> Stride </td> 
 <td> Padding </td> 
 </tr> 
 <tr> 
 <td> 24*24*3 </td> 
 <td> conv1 relu1 </td> 
 <td> 3*3*28 </td> 
 <td> 1 </td> 
 <td>Same</td> 
 </tr> 
 <tr> 
 <td> 24*24*28 </td> 
 <td> pool1_conv1 pool1_relu1 </td> 
 <td>3*3*28</td> 
 <td>2</td> 
 <td>Same</td> 
 </tr> 
 <tr> 
 <td>12*12*28</td> 
 <td> conv2 relu2 </td> 
 <td>3*3*48</td> 
 <td>1</td> 
 <td>Same</td> 
 </tr> 
 <tr> 
 <td>12*12*48</td> 
 <td> pool2_conv3 poo2_relu3 </td> 
 <td>3*3*48</td> 
 <td>2</td> 
 <td>Same</td> 
 </tr> 
 <tr> 
 <td>6*6*48</td> 
 <td> conv3 relu3 </td> 
 <td>3*3*64</td> 
 <td>2</td> 
 <td>Same</td> 
 </tr> 
 <tr> 
 <td>3*3*64</td> 
 <td>&nbsp;</td> 
 <td> &nbsp; </td> 
 <td> &nbsp; </td> 
 <td>&nbsp;</td> 
 </tr> 
 </tbody> 
</table>

## Onet ##

<table> 
 <tbody> 
 <tr> 
 <td> Feature size </td> 
 <td>name</td> 
 <td> Kernel size </td> 
 <td> Stride </td> 
 <td> Padding </td> 
 </tr> 
 <tr> 
 <td> 48*48*3 </td> 
 <td> conv1 relu1 </td> 
 <td> 3*3*32 </td> 
 <td> 1 </td> 
 <td> Same </td> 
 </tr> 
 <tr> 
 <td>48*48*32</td> 
 <td> conv2 relu2 </td> 
 <td>3*3*32</td> 
 <td> 2 </td> 
 <td> Same </td> 
 </tr> 
 <tr> 
 <td> 24*24*32 </td> 
 <td> conv3 relu3 </td> 
 <td>3*3*64</td> 
 <td> 1 </td> 
 <td> Same </td> 
 </tr> 
 <tr> 
 <td> 24*24*64 </td> 
 <td> conv4_ relu4_ </td> 
 <td>3*3*64</td> 
 <td> 2 </td> 
 <td> Same </td> 
 </tr> 
 <tr> 
 <td> 12*12*64 </td> 
 <td> conv5_ relu5_ </td> 
 <td>3*3*128</td> 
 <td> 2 </td> 
 <td> Same </td> 
 </tr> 
 <tr> 
 <td> 6*6*128 </td> 
 <td> conv6_ relu6_ </td> 
 <td>3*3*128</td> 
 <td> 2 </td> 
 <td> Same </td> 
 </tr> 
 <tr> 
 <td>3*3*128</td> 
 <td>&nbsp;</td> 
 <td>&nbsp;</td> 
 <td>&nbsp;</td> 
 <td>&nbsp;</td> 
 </tr> 
 </tbody> 
</table>

# 三、c代码端结构的更改 #

## 3.1 函数的定义 ##

在network.cpp之中

void relu(struct pBox *pbox, mydataFmt *pbias){
 if (pbox->pdata == NULL){
 cout << "the Relu feature is NULL!!" << endl;
 return;
 }
 if (pbias == NULL){
 cout << "the Relu bias is NULL!!" << endl;
 return;
 }
 mydataFmt *op = pbox->pdata;
 mydataFmt *pb = pbias;
 
 long dis = pbox->width*pbox->height;
 for(int channel =0;channel<pbox->channel; channel++){
 for(int col=0; col<dis; col++){
 *op += *pb;
 if(*op<0)*op=0;
 op++;
 }
 pb++;
 }
 }
 void prelu(struct pBox *pbox, mydataFmt *pbias, mydataFmt *prelu_gmma){
 if (pbox->pdata == NULL){
 cout << "the Relu feature is NULL!!" << endl;
 return;
 }
 if (pbias == NULL){
 cout << "the Relu bias is NULL!!" << endl;
 return;
 }
 mydataFmt *op = pbox->pdata;
 mydataFmt *pb = pbias;
 mydataFmt *pg = prelu_gmma;
 
 long dis = pbox->width*pbox->height;
 for(int channel =0;channel<pbox->channel; channel++){
 for(int col=0; col<dis; col++){
 *op = *op + *pb;
 *op = (*op>0)?(*op):((*op)*(*pg));
 op++;
 }
 pb++;
 pg++;
 }
 }

## 3.2 网络的更改 ##

因为relu直接将值进行了替换，所以不存在开辟内存空间的问题，只有运算，并且维度一样，可以直接将prelu函数换位relu函数。

[https_blog.csdn.net_weixin_36474809_article_details_82856171]: https://blog.csdn.net/weixin_36474809/article/details/82856171
[https_blog.csdn.net_weixin_36474809_article_details_82752199]: https://blog.csdn.net/weixin_36474809/article/details/82752199
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl8zNjQ3NDgwOQ_size_16_color_FFFFFF_t_70]: /images/20220413/74b2f8ecc2df446d9258bd64322f6f83.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl8zNjQ3NDgwOQ_size_16_color_FFFFFF_t_70 1]: /images/20220413/3da79bbe5b684ded8fc76f0b78f287e5.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl8zNjQ3NDgwOQ_size_16_color_FFFFFF_t_70 2]: /images/20220413/3b06bc61601948cd94228d84341383cd.png