【nn.Parameter】Pytorch特征融合自适应权重设置（可学习权重使用）

最新推荐文章于 2024-05-24 18:40:42 发布
最新推荐文章于 2024-05-24 18:40:42 发布
阅读量2.5k 收藏 13
点赞数 1
分类专栏： 机器视觉 文章标签： pytorch 学习 深度学习
原文链接：https://blog.csdn.net/weixin_43312117/article/details/121374486
版权

2021年11月17日11:32:14
今天我们来完成Pytorch自适应可学习权重系数，在进行特征融合时，给不同特征图分配可学习的权重！

原文：基于自适应特征融合与转换的小样本图像分类（2021）
期刊：计算机工程与应用（中文核心、CSCD扩展版）

实现这篇论文里面多特征融合的分支！

实现自适应特征处理模块如下图所示：
在这里插入图片描述
特征融合公式如下：

          F
         
         
          
           f
          
          
           f
          
         
        
        
         =
        
        
         
          α
         
         
          1
         
        
        
         ∗
        
        
         
          F
         
         
          
           i
          
          
           d
          
         
        
        
         +
        
        
         
          α
         
         
          2
         
        
        
         ∗
        
        
         
          F
         
         
          
           d
          
          
           c
          
          
           o
          
          
           n
          
          
           v
          
         
        
        
         +
        
        
         
          α
         
         
          3
         
        
        
         ∗
        
        
         
          F
         
         
          max
         
         
          ⁡
         
        
        
         +
        
        
         
          α
         
         
          4
         
        
        
         ∗
        
        
         
          F
         
         
          
           a
          
          
           v
          
          
           g
          
         
        
       
      
     
    
    
     
      
       
      
     
     
      
       
        
        
         
          a
         
         
          i
         
        
        
         =
        
        
         
          
           e
          
          
           
            w
           
           
            i
           
          
         
         
          
           
            Σ
           
           
            j
           
          
          
           
            e
           
           
            
             w
            
            
             j
            
           
          
         
        
        
         (
        
        
         i
        
        
         =
        
        
         1
        
        
         ,
        
        
         2
        
        
         ,
        
        
         3
        
        
         ,
        
        
         4
        
        
         ;
        
        
         j
        
        
         =
        
        
         1
        
        
         ,
        
        
         2
        
        
         ,
        
        
         3
        
        
         ,
        
        
         4
        
        
         )
        
       
      
     
    
   
   
     <span class="MathJax_Preview" style="color: inherit; display: none;"></span><div class="MathJax_Display"><span class="MathJax MathJax_FullWidth" id="MathJax-Element-1-Frame" tabindex="0" data-mathml="<math xmlns=&quot;http://www.w3.org/1998/Math/MathML&quot; display=&quot;block&quot;><mtable columnalign=&quot;right left right left right left right left right left right left&quot; rowspacing=&quot;3pt&quot; columnspacing=&quot;0em 2em 0em 2em 0em 2em 0em 2em 0em 2em 0em&quot; displaystyle=&quot;true&quot;><mtr><mtd /><mtd><msub><mi>F</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>f</mi><mi>f</mi></mrow></msub><mo>=</mo><msub><mi>&amp;#x03B1;</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mn>1</mn></mrow></msub><mo>&amp;#x2217;</mo><msub><mi>F</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>i</mi><mi>d</mi></mrow></msub><mo>+</mo><msub><mi>&amp;#x03B1;</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mn>2</mn></mrow></msub><mo>&amp;#x2217;</mo><msub><mi>F</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>d</mi><mi>c</mi><mi>o</mi><mi>n</mi><mi>v</mi></mrow></msub><mo>+</mo><msub><mi>&amp;#x03B1;</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mn>3</mn></mrow></msub><mo>&amp;#x2217;</mo><msub><mi>F</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mo movablelimits=&quot;true&quot; form=&quot;prefix&quot;>max</mo></mrow></msub><mo>+</mo><msub><mi>&amp;#x03B1;</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mn>4</mn></mrow></msub><mo>&amp;#x2217;</mo><msub><mi>F</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>a</mi><mi>v</mi><mi>g</mi></mrow></msub></mtd></mtr><mtr><mtd /><mtd><msub><mi>a</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>i</mi></mrow></msub><mo>=</mo><mfrac><msup><mi>e</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><msub><mi>w</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>i</mi></mrow></msub></mrow></msup><mrow><msub><mi mathvariant=&quot;normal&quot;>&amp;#x03A3;</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>j</mi></mrow></msub><msup><mi>e</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><msub><mi>w</mi><mrow class=&quot;MJX-TeXAtom-ORD&quot;><mi>j</mi></mrow></msub></mrow></msup></mrow></mfrac><mo stretchy=&quot;false&quot;>(</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>,</mo><mn>3</mn><mo>,</mo><mn>4</mn><mo>;</mo><mi>j</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>,</mo><mn>3</mn><mo>,</mo><mn>4</mn><mo stretchy=&quot;false&quot;>)</mo></mtd></mtr></mtable></math>" role="presentation" style="position: relative;"><nobr aria-hidden="true"><span class="math" id="MathJax-Span-1" style="width: 100%; display: inline-block; min-width: 22.53em;"><span style="display: inline-block; position: relative; width: 100%; height: 0px; font-size: 102%;"><span style="position: absolute; clip: rect(2.432em, 1022.07em, 6.482em, -999.997em); top: -4.705em; left: 0em; width: 100%;"><span class="mrow" id="MathJax-Span-2"><span style="display: inline-block; position: relative; width: 100%; height: 0px;"><span style="position: absolute; clip: rect(2.432em, 1022.07em, 6.482em, -999.997em); top: -4.705em; left: 50%; margin-left: -11.033em;"><span class="mtable" id="MathJax-Span-3" style="padding-left: 0.154em;"><span style="display: inline-block; position: relative; width: 21.922em; height: 0px;"><span style="position: absolute; clip: rect(2.534em, 1000em, 4.812em, -999.997em); top: -3.997em; left: 0em;"><span style="display: inline-block; position: relative; width: 0em; height: 0px;"><span style="position: absolute; clip: rect(3.85em, 1000em, 4.154em, -999.997em); top: -5.313em; right: 0em;"><span class="mtd" id="MathJax-Span-4"><span class="mrow" id="MathJax-Span-5"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; clip: rect(3.85em, 1000em, 4.154em, -999.997em); top: -3.339em; right: 0em;"><span class="mtd" id="MathJax-Span-69"><span class="mrow" id="MathJax-Span-70"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; clip: rect(2.483em, 1021.92em, 6.381em, -999.997em); top: -4.604em; left: 0em;"><span style="display: inline-block; position: relative; width: 21.922em; height: 0px;"><span style="position: absolute; width: 100%; clip: rect(3.192em, 1021.92em, 4.457em, -999.997em); top: -5.313em; left: 0em;"><span class="mtd" id="MathJax-Span-6"><span class="mrow" id="MathJax-Span-7"><span style="display: inline-block; position: relative; width: 100%; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1021.92em, 4.457em, -999.997em); top: -3.997em; left: 50%; margin-left: -10.932em;"><span class="msubsup" id="MathJax-Span-8"><span style="display: inline-block; position: relative; width: 1.521em; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1000.76em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-9" style="font-family: MathJax_Math-italic;">F<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.104em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-10"><span class="mrow" id="MathJax-Span-11"><span style="display: inline-block; position: relative; width: 0.762em; height: 0px;"><span style="position: absolute; clip: rect(3.344em, 1000.76em, 4.305em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-12" style="font-size: 70.7%; font-family: MathJax_Math-italic;">f<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.053em;"></span></span><span class="mi" id="MathJax-Span-13" style="font-size: 70.7%; font-family: MathJax_Math-italic;">f<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.053em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-14" style="font-family: MathJax_Main; padding-left: 0.256em;">=</span><span class="msubsup" id="MathJax-Span-15" style="padding-left: 0.256em;"><span style="display: inline-block; position: relative; width: 1.066em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.61em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-16" style="font-family: MathJax_Math-italic;">α</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-17"><span class="mrow" id="MathJax-Span-18"><span style="display: inline-block; position: relative; width: 0.357em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.31em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mn" id="MathJax-Span-19" style="font-size: 70.7%; font-family: MathJax_Main;">1</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-20" style="font-family: MathJax_Main; padding-left: 0.205em;">∗</span><span class="msubsup" id="MathJax-Span-21" style="padding-left: 0.205em;"><span style="display: inline-block; position: relative; width: 1.319em; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1000.76em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-22" style="font-family: MathJax_Math-italic;">F<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.104em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-23"><span class="mrow" id="MathJax-Span-24"><span style="display: inline-block; position: relative; width: 0.61em; height: 0px;"><span style="position: absolute; clip: rect(3.344em, 1000.61em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-25" style="font-size: 70.7%; font-family: MathJax_Math-italic;">i</span><span class="mi" id="MathJax-Span-26" style="font-size: 70.7%; font-family: MathJax_Math-italic;">d<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.003em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-27" style="font-family: MathJax_Main; padding-left: 0.205em;">+</span><span class="msubsup" id="MathJax-Span-28" style="padding-left: 0.205em;"><span style="display: inline-block; position: relative; width: 1.066em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.61em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-29" style="font-family: MathJax_Math-italic;">α</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-30"><span class="mrow" id="MathJax-Span-31"><span style="display: inline-block; position: relative; width: 0.357em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.31em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mn" id="MathJax-Span-32" style="font-size: 70.7%; font-family: MathJax_Main;">2</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-33" style="font-family: MathJax_Main; padding-left: 0.205em;">∗</span><span class="msubsup" id="MathJax-Span-34" style="padding-left: 0.205em;"><span style="display: inline-block; position: relative; width: 2.483em; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1000.76em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-35" style="font-family: MathJax_Math-italic;">F<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.104em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-36"><span class="mrow" id="MathJax-Span-37"><span style="display: inline-block; position: relative; width: 1.774em; height: 0px;"><span style="position: absolute; clip: rect(3.344em, 1001.77em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-38" style="font-size: 70.7%; font-family: MathJax_Math-italic;">d<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.003em;"></span></span><span class="mi" id="MathJax-Span-39" style="font-size: 70.7%; font-family: MathJax_Math-italic;">c</span><span class="mi" id="MathJax-Span-40" style="font-size: 70.7%; font-family: MathJax_Math-italic;">o</span><span class="mi" id="MathJax-Span-41" style="font-size: 70.7%; font-family: MathJax_Math-italic;">n</span><span class="mi" id="MathJax-Span-42" style="font-size: 70.7%; font-family: MathJax_Math-italic;">v</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-43" style="font-family: MathJax_Main; padding-left: 0.205em;">+</span><span class="msubsup" id="MathJax-Span-44" style="padding-left: 0.205em;"><span style="display: inline-block; position: relative; width: 1.066em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.61em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-45" style="font-family: MathJax_Math-italic;">α</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-46"><span class="mrow" id="MathJax-Span-47"><span style="display: inline-block; position: relative; width: 0.357em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.31em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mn" id="MathJax-Span-48" style="font-size: 70.7%; font-family: MathJax_Main;">3</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-49" style="font-family: MathJax_Main; padding-left: 0.205em;">∗</span><span class="msubsup" id="MathJax-Span-50" style="padding-left: 0.205em;"><span style="display: inline-block; position: relative; width: 2.027em; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1000.76em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-51" style="font-family: MathJax_Math-italic;">F<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.104em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-52"><span class="mrow" id="MathJax-Span-53"><span style="display: inline-block; position: relative; width: 1.319em; height: 0px;"><span style="position: absolute; clip: rect(3.546em, 1001.32em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mo" id="MathJax-Span-54" style="font-size: 70.7%; font-family: MathJax_Main;">max</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-55" style="font-family: MathJax_Main; padding-left: 0.205em;">+</span><span class="msubsup" id="MathJax-Span-56" style="padding-left: 0.205em;"><span style="display: inline-block; position: relative; width: 1.066em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.61em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-57" style="font-family: MathJax_Math-italic;">α</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-58"><span class="mrow" id="MathJax-Span-59"><span style="display: inline-block; position: relative; width: 0.357em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.36em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mn" id="MathJax-Span-60" style="font-size: 70.7%; font-family: MathJax_Main;">4</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-61" style="font-family: MathJax_Main; padding-left: 0.205em;">∗</span><span class="msubsup" id="MathJax-Span-62" style="padding-left: 0.205em;"><span style="display: inline-block; position: relative; width: 1.774em; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1000.76em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-63" style="font-family: MathJax_Math-italic;">F<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.104em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.661em;"><span class="texatom" id="MathJax-Span-64"><span class="mrow" id="MathJax-Span-65"><span style="display: inline-block; position: relative; width: 1.066em; height: 0px;"><span style="position: absolute; clip: rect(3.546em, 1001.07em, 4.305em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-66" style="font-size: 70.7%; font-family: MathJax_Math-italic;">a</span><span class="mi" id="MathJax-Span-67" style="font-size: 70.7%; font-family: MathJax_Math-italic;">v</span><span class="mi" id="MathJax-Span-68" style="font-size: 70.7%; font-family: MathJax_Math-italic;">g<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.003em;"></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; width: 100%; clip: rect(2.483em, 1015.95em, 5.115em, -999.997em); top: -3.339em; left: 0em;"><span class="mtd" id="MathJax-Span-71"><span class="mrow" id="MathJax-Span-72"><span style="display: inline-block; position: relative; width: 100%; height: 0px;"><span style="position: absolute; clip: rect(2.483em, 1015.95em, 5.115em, -999.997em); top: -3.997em; left: 50%; margin-left: -7.996em;"><span class="msubsup" id="MathJax-Span-73"><span style="display: inline-block; position: relative; width: 0.863em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.51em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-74" style="font-family: MathJax_Math-italic;">a</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.509em;"><span class="texatom" id="MathJax-Span-75"><span class="mrow" id="MathJax-Span-76"><span style="display: inline-block; position: relative; width: 0.256em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.21em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-77" style="font-size: 70.7%; font-family: MathJax_Math-italic;">i</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="mo" id="MathJax-Span-78" style="font-family: MathJax_Main; padding-left: 0.256em;">=</span><span class="mfrac" id="MathJax-Span-79" style="padding-left: 0.256em;"><span style="display: inline-block; position: relative; width: 2.534em; height: 0px; margin-right: 0.104em; margin-left: 0.104em;"><span style="position: absolute; clip: rect(3.192em, 1001.27em, 4.154em, -999.997em); top: -4.655em; left: 50%; margin-left: -0.656em;"><span class="msubsup" id="MathJax-Span-80"><span style="display: inline-block; position: relative; width: 1.268em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.41em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-81" style="font-family: MathJax_Math-italic;">e</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -4.351em; left: 0.458em;"><span class="texatom" id="MathJax-Span-82"><span class="mrow" id="MathJax-Span-83"><span style="display: inline-block; position: relative; width: 0.711em; height: 0px;"><span style="position: absolute; clip: rect(3.546em, 1000.71em, 4.255em, -999.997em); top: -3.997em; left: 0em;"><span class="msubsup" id="MathJax-Span-84"><span style="display: inline-block; position: relative; width: 0.711em; height: 0px;"><span style="position: absolute; clip: rect(3.546em, 1000.51em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-85" style="font-size: 70.7%; font-family: MathJax_Math-italic;">w</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.895em; left: 0.509em;"><span class="texatom" id="MathJax-Span-86"><span class="mrow" id="MathJax-Span-87"><span style="display: inline-block; position: relative; width: 0.154em; height: 0px;"><span style="position: absolute; clip: rect(3.495em, 1000.15em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-88" style="font-size: 50%; font-family: MathJax_Math-italic;">i</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; clip: rect(3.192em, 1002.38em, 4.457em, -999.997em); top: -3.288em; left: 50%; margin-left: -1.212em;"><span class="mrow" id="MathJax-Span-89"><span style="display: inline-block; position: relative; width: 2.382em; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1002.38em, 4.457em, -999.997em); top: -3.997em; left: 0em;"><span class="msubsup" id="MathJax-Span-90"><span style="display: inline-block; position: relative; width: 1.066em; height: 0px;"><span style="position: absolute; clip: rect(3.192em, 1000.66em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-91" style="font-family: MathJax_Main;">Σ</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.845em; left: 0.711em;"><span class="texatom" id="MathJax-Span-92"><span class="mrow" id="MathJax-Span-93"><span style="display: inline-block; position: relative; width: 0.306em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.31em, 4.305em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-94" style="font-size: 70.7%; font-family: MathJax_Math-italic;">j</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span class="msubsup" id="MathJax-Span-95"><span style="display: inline-block; position: relative; width: 1.319em; height: 0px;"><span style="position: absolute; clip: rect(3.394em, 1000.41em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-96" style="font-family: MathJax_Math-italic;">e</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -4.3em; left: 0.458em;"><span class="texatom" id="MathJax-Span-97"><span class="mrow" id="MathJax-Span-98"><span style="display: inline-block; position: relative; width: 0.762em; height: 0px;"><span style="position: absolute; clip: rect(3.546em, 1000.76em, 4.356em, -999.997em); top: -3.997em; left: 0em;"><span class="msubsup" id="MathJax-Span-99"><span style="display: inline-block; position: relative; width: 0.762em; height: 0px;"><span style="position: absolute; clip: rect(3.546em, 1000.51em, 4.154em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-100" style="font-size: 70.7%; font-family: MathJax_Math-italic;">w</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; top: -3.895em; left: 0.509em;"><span class="texatom" id="MathJax-Span-101"><span class="mrow" id="MathJax-Span-102"><span style="display: inline-block; position: relative; width: 0.205em; height: 0px;"><span style="position: absolute; clip: rect(3.495em, 1000.21em, 4.255em, -999.997em); top: -3.997em; left: 0em;"><span class="mi" id="MathJax-Span-103" style="font-size: 50%; font-family: MathJax_Math-italic;">j</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span><span style="position: absolute; clip: rect(0.863em, 1002.53em, 1.217em, -999.997em); top: -1.263em; left: 0em;"><span style="display: inline-block; overflow: hidden; vertical-align: 0em; border-top: 1.3px solid; width: 2.534em; height: 0px;"></span><span style="display: inline-block; width: 0px; height: 1.066em;"></span></span></span></span><span class="mo" id="MathJax-Span-104" style="font-family: MathJax_Main;">(</span><span class="mi" id="MathJax-Span-105" style="font-family: MathJax_Math-italic;">i</span><span class="mo" id="MathJax-Span-106" style="font-family: MathJax_Main; padding-left: 0.256em;">=</span><span class="mn" id="MathJax-Span-107" style="font-family: MathJax_Main; padding-left: 0.256em;">1</span><span class="mo" id="MathJax-Span-108" style="font-family: MathJax_Main;">,</span><span class="mn" id="MathJax-Span-109" style="font-family: MathJax_Main; padding-left: 0.154em;">2</span><span class="mo" id="MathJax-Span-110" style="font-family: MathJax_Main;">,</span><span class="mn" id="MathJax-Span-111" style="font-family: MathJax_Main; padding-left: 0.154em;">3</span><span class="mo" id="MathJax-Span-112" style="font-family: MathJax_Main;">,</span><span class="mn" id="MathJax-Span-113" style="font-family: MathJax_Main; padding-left: 0.154em;">4</span><span class="mo" id="MathJax-Span-114" style="font-family: MathJax_Main;">;</span><span class="mi" id="MathJax-Span-115" style="font-family: MathJax_Math-italic; padding-left: 0.154em;">j</span><span class="mo" id="MathJax-Span-116" style="font-family: MathJax_Main; padding-left: 0.256em;">=</span><span class="mn" id="MathJax-Span-117" style="font-family: MathJax_Main; padding-left: 0.256em;">1</span><span class="mo" id="MathJax-Span-118" style="font-family: MathJax_Main;">,</span><span class="mn" id="MathJax-Span-119" style="font-family: MathJax_Main; padding-left: 0.154em;">2</span><span class="mo" id="MathJax-Span-120" style="font-family: MathJax_Main;">,</span><span class="mn" id="MathJax-Span-121" style="font-family: MathJax_Main; padding-left: 0.154em;">3</span><span class="mo" id="MathJax-Span-122" style="font-family: MathJax_Main;">,</span><span class="mn" id="MathJax-Span-123" style="font-family: MathJax_Main; padding-left: 0.154em;">4</span><span class="mo" id="MathJax-Span-124" style="font-family: MathJax_Main;">)</span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.002em;"></span></span></span><span style="display: inline-block; width: 0px; height: 4.609em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.71em;"></span></span></span></span><span style="display: inline-block; width: 0px; height: 4.71em;"></span></span></span><span style="display: inline-block; overflow: hidden; vertical-align: -1.701em; border-left: 0px solid; width: 0px; height: 3.927em;"></span></span></nobr><span class="MJX_Assistive_MathML MJX_Assistive_MathML_Block" role="presentation"><math xmlns="http://www.w3.org/1998/Math/MathML" display="block"><mtable columnalign="right left right left right left right left right left right left" rowspacing="3pt" columnspacing="0em 2em 0em 2em 0em 2em 0em 2em 0em 2em 0em" displaystyle="true"><mtr><mtd></mtd><mtd><msub><mi>F</mi><mrow class="MJX-TeXAtom-ORD"><mi>f</mi><mi>f</mi></mrow></msub><mo>=</mo><msub><mi>α</mi><mrow class="MJX-TeXAtom-ORD"><mn>1</mn></mrow></msub><mo>∗</mo><msub><mi>F</mi><mrow class="MJX-TeXAtom-ORD"><mi>i</mi><mi>d</mi></mrow></msub><mo>+</mo><msub><mi>α</mi><mrow class="MJX-TeXAtom-ORD"><mn>2</mn></mrow></msub><mo>∗</mo><msub><mi>F</mi><mrow class="MJX-TeXAtom-ORD"><mi>d</mi><mi>c</mi><mi>o</mi><mi>n</mi><mi>v</mi></mrow></msub><mo>+</mo><msub><mi>α</mi><mrow class="MJX-TeXAtom-ORD"><mn>3</mn></mrow></msub><mo>∗</mo><msub><mi>F</mi><mrow class="MJX-TeXAtom-ORD"><mo movablelimits="true" form="prefix">max</mo></mrow></msub><mo>+</mo><msub><mi>α</mi><mrow class="MJX-TeXAtom-ORD"><mn>4</mn></mrow></msub><mo>∗</mo><msub><mi>F</mi><mrow class="MJX-TeXAtom-ORD"><mi>a</mi><mi>v</mi><mi>g</mi></mrow></msub></mtd></mtr><mtr><mtd></mtd><mtd><msub><mi>a</mi><mrow class="MJX-TeXAtom-ORD"><mi>i</mi></mrow></msub><mo>=</mo><mfrac><msup><mi>e</mi><mrow class="MJX-TeXAtom-ORD"><msub><mi>w</mi><mrow class="MJX-TeXAtom-ORD"><mi>i</mi></mrow></msub></mrow></msup><mrow><msub><mi mathvariant="normal">Σ</mi><mrow class="MJX-TeXAtom-ORD"><mi>j</mi></mrow></msub><msup><mi>e</mi><mrow class="MJX-TeXAtom-ORD"><msub><mi>w</mi><mrow class="MJX-TeXAtom-ORD"><mi>j</mi></mrow></msub></mrow></msup></mrow></mfrac><mo stretchy="false">(</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>,</mo><mn>3</mn><mo>,</mo><mn>4</mn><mo>;</mo><mi>j</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>,</mo><mn>3</mn><mo>,</mo><mn>4</mn><mo stretchy="false">)</mo></mtd></mtr></mtable></math></span></span></div><script type="math/tex; mode=display" id="MathJax-Element-1">\begin{aligned} &F_{f f}=\alpha_{1} * F_{i d}+\alpha_{2} * F_{dconv}+\alpha_{3} * F_{\max }+\alpha_{4} * F_{a v g} \\ &a_{i}=\frac{e^{w_{i}}}{\Sigma_{j} e^{w_{j}}}(i=1,2,3,4 ; j=1,2,3,4)\end{aligned}</script> 
   
  
 </span><span class="katex-html"><span class="base"><span class="strut" style="height: 4.1135em; vertical-align: -1.80675em;"></span><span class="mord"><span class="mtable"><span class="col-align-r"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 2.30675em;"><span class="" style="top: -4.80814em;"><span class="pstrut" style="height: 3.34139em;"></span><span class="mord"></span></span><span class="" style="top: -2.80675em;"><span class="pstrut" style="height: 3.34139em;"></span><span class="mord"></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 1.80675em;"><span class=""></span></span></span></span></span><span class="col-align-l"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 2.30675em;"><span class="" style="top: -4.80814em;"><span class="pstrut" style="height: 3.34139em;"></span><span class="mord"><span class="mord"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.13889em;">F</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.336108em;"><span class="" style="top: -2.55em; margin-left: -0.13889em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathdefault mtight" style="margin-right: 0.10764em;">f</span><span class="mord mathdefault mtight" style="margin-right: 0.10764em;">f</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.286108em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mrel">=</span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.0037em;">α</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.301108em;"><span class="" style="top: -2.55em; margin-left: -0.0037em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">1</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mbin">∗</span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.13889em;">F</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.336108em;"><span class="" style="top: -2.55em; margin-left: -0.13889em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathdefault mtight">i</span><span class="mord mathdefault mtight">d</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mbin">+</span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.0037em;">α</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.301108em;"><span class="" style="top: -2.55em; margin-left: -0.0037em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">2</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mbin">∗</span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.13889em;">F</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.336108em;"><span class="" style="top: -2.55em; margin-left: -0.13889em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathdefault mtight">d</span><span class="mord mathdefault mtight">c</span><span class="mord mathdefault mtight">o</span><span class="mord mathdefault mtight">n</span><span class="mord mathdefault mtight" style="margin-right: 0.03588em;">v</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mbin">+</span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.0037em;">α</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.301108em;"><span class="" style="top: -2.55em; margin-left: -0.0037em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">3</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mbin">∗</span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.13889em;">F</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.151392em;"><span class="" style="top: -2.15139em; margin-left: -0.13889em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.30139em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mop mtight">max</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mbin">+</span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.0037em;">α</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.301108em;"><span class="" style="top: -2.55em; margin-left: -0.0037em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">4</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mbin">∗</span><span class="mspace" style="margin-right: 0.222222em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.13889em;">F</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.151392em;"><span class="" style="top: -2.55em; margin-left: -0.13889em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathdefault mtight">a</span><span class="mord mathdefault mtight" style="margin-right: 0.03588em;">v</span><span class="mord mathdefault mtight" style="margin-right: 0.03588em;">g</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.286108em;"><span class=""></span></span></span></span></span></span></span></span><span class="" style="top: -2.80675em;"><span class="pstrut" style="height: 3.34139em;"></span><span class="mord"><span class="mord"></span><span class="mord"><span class="mord mathdefault">a</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.311664em;"><span class="" style="top: -2.55em; margin-left: 0em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathdefault mtight">i</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mrel">=</span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mord"><span class="mopen nulldelimiter"></span><span class="mfrac"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 1.34139em;"><span class="" style="top: -2.314em;"><span class="pstrut" style="height: 3em;"></span><span class="mord"><span class="mord"><span class="mord">Σ</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.311664em;"><span class="" style="top: -2.55em; margin-left: 0em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathdefault mtight" style="margin-right: 0.05724em;">j</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.286108em;"><span class=""></span></span></span></span></span></span><span class="mord"><span class="mord mathdefault">e</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height: 0.606462em;"><span class="" style="top: -3.00507em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathdefault mtight" style="margin-right: 0.02691em;">w</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.328086em;"><span class="" style="top: -2.357em; margin-left: -0.02691em; margin-right: 0.0714286em;"><span class="pstrut" style="height: 2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mathdefault mtight" style="margin-right: 0.05724em;">j</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.281886em;"><span class=""></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span><span class="" style="top: -3.23em;"><span class="pstrut" style="height: 3em;"></span><span class="frac-line" style="border-bottom-width: 0.04em;"></span></span><span class="" style="top: -3.677em;"><span class="pstrut" style="height: 3em;"></span><span class="mord"><span class="mord"><span class="mord mathdefault">e</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height: 0.664392em;"><span class="" style="top: -3.063em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathdefault mtight" style="margin-right: 0.02691em;">w</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.328086em;"><span class="" style="top: -2.357em; margin-left: -0.02691em; margin-right: 0.0714286em;"><span class="pstrut" style="height: 2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mathdefault mtight">i</span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.143em;"><span class=""></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.972108em;"><span class=""></span></span></span></span></span><span class="mclose nulldelimiter"></span></span><span class="mopen">(</span><span class="mord mathdefault">i</span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mrel">=</span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mord">1</span><span class="mpunct">,</span><span class="mspace" style="margin-right: 0.166667em;"></span><span class="mord">2</span><span class="mpunct">,</span><span class="mspace" style="margin-right: 0.166667em;"></span><span class="mord">3</span><span class="mpunct">,</span><span class="mspace" style="margin-right: 0.166667em;"></span><span class="mord">4</span><span class="mpunct">;</span><span class="mspace" style="margin-right: 0.166667em;"></span><span class="mord mathdefault" style="margin-right: 0.05724em;">j</span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mrel">=</span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mord">1</span><span class="mpunct">,</span><span class="mspace" style="margin-right: 0.166667em;"></span><span class="mord">2</span><span class="mpunct">,</span><span class="mspace" style="margin-right: 0.166667em;"></span><span class="mord">3</span><span class="mpunct">,</span><span class="mspace" style="margin-right: 0.166667em;"></span><span class="mord">4</span><span class="mclose">)</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 1.80675em;"><span class=""></span></span></span></span></span></span></span></span></span></span></span></span><br> 其中，<span class="katex--inline"><span class="katex"><span class="katex-mathml">

 
  
   
    
     α
    
    
     i
    
   
  
  
   \alpha_i
  
 
</span><span class="katex-html"><span class="base"><span class="strut" style="height: 0.58056em; vertical-align: -0.15em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.0037em;">α</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.311664em;"><span class="" style="top: -2.55em; margin-left: -0.0037em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathdefault mtight">i</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span></span></span></span></span>为归一化权重，<span class="katex--inline"><span class="katex"><span class="katex-mathml">

 
  
   
    Σ
   
   
    
     α
    
    
     i
    
   
   
    =
   
   
    1
   
  
  
   \Sigma\alpha_i=1
  
 
</span><span class="katex-html"><span class="base"><span class="strut" style="height: 0.83333em; vertical-align: -0.15em;"></span><span class="mord">Σ</span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.0037em;">α</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.311664em;"><span class="" style="top: -2.55em; margin-left: -0.0037em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathdefault mtight">i</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span><span class="mspace" style="margin-right: 0.277778em;"></span><span class="mrel">=</span><span class="mspace" style="margin-right: 0.277778em;"></span></span><span class="base"><span class="strut" style="height: 0.64444em; vertical-align: 0em;"></span><span class="mord">1</span></span></span></span></span>，<span class="katex--inline"><span class="katex"><span class="katex-mathml">

 
  
   
    
     w
    
    
     i
    
   
  
  
   w_i
  
 
</span><span class="katex-html"><span class="base"><span class="strut" style="height: 0.58056em; vertical-align: -0.15em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.02691em;">w</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.311664em;"><span class="" style="top: -2.55em; margin-left: -0.02691em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathdefault mtight">i</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span></span></span></span></span>为初始化权重系数。</p>

结构分析：

  1. 对于一个输入的特征图，有四个分支
  2. 从上往下，第一个分支用的是Maxpooling进行最大池化提取局部特征
  3. 第二个分支用的是Avgpooling进行平均池化提取全局特征
  4. 第三个分支，原文中讲的是“用两组1×1卷积将特征的通道减半压缩，一是为了减少参数量防止过拟合，二是方便后续进行卷积特征拼接进行加性融合；接着在第一组1×1卷积后加入两组3×3卷积来替代5×5卷积后按通道进行拼接（Combine按通道拼接）。”原文将这个分支称作双卷积分支DConv，卷积能提取丰富特征，在拼接后接入一个SE注意力模块。
  5. 第三个分支，是残差分支Identity，把输入直接跳跃连接加过去，保留原始特征

模型分析：

  1. 分析下模块结构，既然对于特征融合，最后的操作是Add,那么4个分支输出的特征图大小和维度是相同的！跳跃连接时原图的大小和维度都没有变，所以我们让四个分支的输出和原图大小保持一致
  2. 原文在3.2.2参数设置里面说：最大池化支路池化尺寸设为3，平均池化分支池化尺寸设为2
  3. 初始化各特征的权重全为1，使用nn.Parameter实现
  4. 输入图像的大小为3×84×84

AFP模块Pytorch实现

"""
Author: yida
Time is: 2021/11/17 15:45 
this Code:
1.实现<基于自适应特征融合与转换的小样本图像分类>中的自适应特征处理模块AFP
2.演示: nn.Parameter的使用
"""
import torch
import torch.nn as nn

class AFP(nn.Module):
def init(self):
super(AFP, self).init()

    self<span class="token punctuation">.</span>branch1 <span class="token operator">=</span> nn<span class="token punctuation">.</span>Sequential<span class="token punctuation">(</span>
        nn<span class="token punctuation">.</span>MaxPool2d<span class="token punctuation">(</span><span class="token number">3</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">,</span> padding<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span><span class="token punctuation">,</span>  <span class="token comment"># 1.最大池化分支,原文设置的尺寸大小为3, 未说明stride以及padding, 为与原图大小保持一致, 使用(3, 1, 1)</span>
    <span class="token punctuation">)</span>
    self<span class="token punctuation">.</span>branch2 <span class="token operator">=</span> nn<span class="token punctuation">.</span>Sequential<span class="token punctuation">(</span>
        nn<span class="token punctuation">.</span>AvgPool2d<span class="token punctuation">(</span><span class="token number">3</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">,</span> padding<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span><span class="token punctuation">,</span>  <span class="token comment"># 2.平均池化分支, 原文设置的池化尺寸为2, 未说明stride以及padding, 为与原图大小保持一致, 使用(3, 1, 1)</span>
    <span class="token punctuation">)</span>

    self<span class="token punctuation">.</span>branch3_1 <span class="token operator">=</span> nn<span class="token punctuation">.</span>Sequential<span class="token punctuation">(</span>
        nn<span class="token punctuation">.</span>Conv2d<span class="token punctuation">(</span><span class="token number">3</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">)</span><span class="token punctuation">,</span>
        nn<span class="token punctuation">.</span>Conv2d<span class="token punctuation">(</span><span class="token number">1</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">,</span> <span class="token number">3</span><span class="token punctuation">,</span> padding<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span><span class="token punctuation">,</span>  <span class="token comment"># 3_1分支, 先用1×1卷积压缩通道维数, 然后使用两个3×3卷积进行特征提取, 由于通道数为3//2, 此时输出维度设为1</span>
        nn<span class="token punctuation">.</span>Conv2d<span class="token punctuation">(</span><span class="token number">1</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">,</span> <span class="token number">3</span><span class="token punctuation">,</span> padding<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span><span class="token punctuation">,</span>
    <span class="token punctuation">)</span>

    self<span class="token punctuation">.</span>branch3_2 <span class="token operator">=</span> nn<span class="token punctuation">.</span>Sequential<span class="token punctuation">(</span>
        nn<span class="token punctuation">.</span>Conv2d<span class="token punctuation">(</span><span class="token number">3</span><span class="token punctuation">,</span> <span class="token number">2</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">)</span><span class="token punctuation">,</span>  <span class="token comment"># 3_2分支, 由于1×1卷积压缩通道维数减半, 但是这儿维度为3, 上面用的1, 所以这儿输出维度设为2</span>
        nn<span class="token punctuation">.</span>Conv2d<span class="token punctuation">(</span><span class="token number">2</span><span class="token punctuation">,</span> <span class="token number">2</span><span class="token punctuation">,</span> <span class="token number">3</span><span class="token punctuation">,</span> padding<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span>
    <span class="token punctuation">)</span>
    <span class="token comment"># 注意力机制</span>
    self<span class="token punctuation">.</span>branch_SE <span class="token operator">=</span> SEblock<span class="token punctuation">(</span>channel<span class="token operator">=</span><span class="token number">3</span><span class="token punctuation">)</span>

    <span class="token comment"># 初始化可学习权重系数</span>
    <span class="token comment"># nn.Parameter 初始化的权重, 如果作用到网络中的话, 那么它会被添加到优化器更新的参数中, 优化器更新的时候会纠正Parameter的值, 使得向损失函数最小化的方向优化</span>

    self<span class="token punctuation">.</span>w <span class="token operator">=</span> nn<span class="token punctuation">.</span>Parameter<span class="token punctuation">(</span>torch<span class="token punctuation">.</span>ones<span class="token punctuation">(</span><span class="token number">4</span><span class="token punctuation">)</span><span class="token punctuation">)</span>  <span class="token comment"># 4个分支, 每个分支设置一个自适应学习权重, 初始化为1, nn.Parameter需放入Tensor类型的数据</span>
    <span class="token comment"># self.w = nn.Parameter(torch.Tensor([0.5, 0.25, 0.15, 0.1]), requires_grad=False)  # 设置固定的权重系数, 不用归一化, 直接乘过去</span>

<span class="token keyword">def</span> <span class="token function">forward</span><span class="token punctuation">(</span>self<span class="token punctuation">,</span> x<span class="token punctuation">)</span><span class="token punctuation">:</span>
    b1 <span class="token operator">=</span> self<span class="token punctuation">.</span>branch1<span class="token punctuation">(</span>x<span class="token punctuation">)</span>
    b2 <span class="token operator">=</span> self<span class="token punctuation">.</span>branch2<span class="token punctuation">(</span>x<span class="token punctuation">)</span>

    b3_1 <span class="token operator">=</span> self<span class="token punctuation">.</span>branch3_1<span class="token punctuation">(</span>x<span class="token punctuation">)</span>
    b3_2 <span class="token operator">=</span> self<span class="token punctuation">.</span>branch3_2<span class="token punctuation">(</span>x<span class="token punctuation">)</span>
    b3_Combine <span class="token operator">=</span> torch<span class="token punctuation">.</span>cat<span class="token punctuation">(</span><span class="token punctuation">(</span>b3_1<span class="token punctuation">,</span> b3_2<span class="token punctuation">)</span><span class="token punctuation">,</span> dim<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span>
    b3 <span class="token operator">=</span> self<span class="token punctuation">.</span>branch_SE<span class="token punctuation">(</span>b3_Combine<span class="token punctuation">)</span>

    b4 <span class="token operator">=</span> x

    <span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">"b1:"</span><span class="token punctuation">,</span> b1<span class="token punctuation">.</span>shape<span class="token punctuation">)</span>
    <span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">"b2:"</span><span class="token punctuation">,</span> b2<span class="token punctuation">.</span>shape<span class="token punctuation">)</span>
    <span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">"b3:"</span><span class="token punctuation">,</span> b3<span class="token punctuation">.</span>shape<span class="token punctuation">)</span>
    <span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">"b4:"</span><span class="token punctuation">,</span> b4<span class="token punctuation">.</span>shape<span class="token punctuation">)</span>

    <span class="token comment"># 归一化权重</span>
    w1 <span class="token operator">=</span> torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">[</span><span class="token number">0</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">)</span><span class="token punctuation">)</span>
    w2 <span class="token operator">=</span> torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">)</span><span class="token punctuation">)</span>
    w3 <span class="token operator">=</span> torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">)</span><span class="token punctuation">)</span>
    w4 <span class="token operator">=</span> torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">[</span><span class="token number">3</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>self<span class="token punctuation">.</span>w<span class="token punctuation">)</span><span class="token punctuation">)</span>

    <span class="token comment"># 多特征融合</span>
    x_out <span class="token operator">=</span> b1 <span class="token operator">*</span> w1 <span class="token operator">+</span> b2 <span class="token operator">*</span> w2 <span class="token operator">+</span> b3 <span class="token operator">*</span> w3 <span class="token operator">+</span> b4 <span class="token operator">*</span> w4
    <span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">"特征融合结果:"</span><span class="token punctuation">,</span> x_out<span class="token punctuation">.</span>shape<span class="token punctuation">)</span>
    <span class="token keyword">return</span> x_out

class SEblock(nn.Module): # 注意力机制模块
def init(self, channel, r=0.5): # channel为输入的维度, r为全连接层缩放比例->控制中间层个数
super(SEblock, self).init()
# 全局均值池化
self.global_avg_pool = nn.AdaptiveAvgPool2d(1)
# 全连接层
self.fc = nn.Sequential(
nn.Linear(channel, int(channel r)), # int(channel * r)取整数
nn.ReLU(),
nn.Linear(int(channel r), channel),
nn.Sigmoid(),
)

<span class="token keyword">def</span> <span class="token function">forward</span><span class="token punctuation">(</span>self<span class="token punctuation">,</span> x<span class="token punctuation">)</span><span class="token punctuation">:</span>
    <span class="token comment"># 对x进行分支计算权重, 进行全局均值池化</span>
    branch <span class="token operator">=</span> self<span class="token punctuation">.</span>global_avg_pool<span class="token punctuation">(</span>x<span class="token punctuation">)</span>
    branch <span class="token operator">=</span> branch<span class="token punctuation">.</span>view<span class="token punctuation">(</span>branch<span class="token punctuation">.</span>size<span class="token punctuation">(</span><span class="token number">0</span><span class="token punctuation">)</span><span class="token punctuation">,</span> <span class="token operator">-</span><span class="token number">1</span><span class="token punctuation">)</span>

    <span class="token comment"># 全连接层得到权重</span>
    weight <span class="token operator">=</span> self<span class="token punctuation">.</span>fc<span class="token punctuation">(</span>branch<span class="token punctuation">)</span>

    <span class="token comment"># 将维度为b, c的weight, reshape成b, c, 1, 1 与 输入x 相乘</span>
    h<span class="token punctuation">,</span> w <span class="token operator">=</span> weight<span class="token punctuation">.</span>shape
    weight <span class="token operator">=</span> torch<span class="token punctuation">.</span>reshape<span class="token punctuation">(</span>weight<span class="token punctuation">,</span> <span class="token punctuation">(</span>h<span class="token punctuation">,</span> w<span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">)</span><span class="token punctuation">)</span>

    <span class="token comment"># 乘积获得结果</span>
    scale <span class="token operator">=</span> weight <span class="token operator">*</span> x
    <span class="token keyword">return</span> scale

if name == ‘main’:
model = AFP()
print(model)
inputs = torch.randn(10, 3, 84, 84)
print("输入维度为: ", inputs.shape)
outputs = model(inputs)
print("输出维度为: ", outputs.shape)

<span class="token comment"># 查看nn.Parameter中值的变化, 训练网络时, 更新优化器之后, 可以循环输出, 查看权重变化</span>
<span class="token keyword">for</span> name<span class="token punctuation">,</span> p <span class="token keyword">in</span> model<span class="token punctuation">.</span>named_parameters<span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">:</span>
    <span class="token keyword">if</span> name <span class="token operator">==</span> <span class="token string">'w'</span><span class="token punctuation">:</span>
        <span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">"特征权重: "</span><span class="token punctuation">,</span> name<span class="token punctuation">)</span>
        w0 <span class="token operator">=</span> <span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">[</span><span class="token number">0</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">.</span>item<span class="token punctuation">(</span><span class="token punctuation">)</span>
        w1 <span class="token operator">=</span> <span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">.</span>item<span class="token punctuation">(</span><span class="token punctuation">)</span>
        w2 <span class="token operator">=</span> <span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">.</span>item<span class="token punctuation">(</span><span class="token punctuation">)</span>
        w3 <span class="token operator">=</span> <span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">[</span><span class="token number">3</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token operator">/</span> torch<span class="token punctuation">.</span><span class="token builtin">sum</span><span class="token punctuation">(</span>torch<span class="token punctuation">.</span>exp<span class="token punctuation">(</span>p<span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">.</span>item<span class="token punctuation">(</span><span class="token punctuation">)</span>
        <span class="token keyword">print</span><span class="token punctuation">(</span>w0<span class="token punctuation">,</span> w1<span class="token punctuation">,</span> w2<span class="token punctuation">,</span> w3<span class="token punctuation">)</span>
  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40
  • 41
  • 42
  • 43
  • 44
  • 45
  • 46
  • 47
  • 48
  • 49
  • 50
  • 51
  • 52
  • 53
  • 54
  • 55
  • 56
  • 57
  • 58
  • 59
  • 60
  • 61
  • 62
  • 63
  • 64
  • 65
  • 66
  • 67
  • 68
  • 69
  • 70
  • 71
  • 72
  • 73
  • 74
  • 75
  • 76
  • 77
  • 78
  • 79
  • 80
  • 81
  • 82
  • 83
  • 84
  • 85
  • 86
  • 87
  • 88
  • 89
  • 90
  • 91
  • 92
  • 93
  • 94
  • 95
  • 96
  • 97
  • 98
  • 99
  • 100
  • 101
  • 102
  • 103
  • 104
  • 105
  • 106
  • 107
  • 108
  • 109
  • 110
  • 111
  • 112
  • 113
  • 114
  • 115
  • 116

nn.Parameter:上图特征融合中的权重系数 

      w
     
     
      i
     
    
   
   
    w_i
   
  
 </span><span class="katex-html"><span class="base"><span class="strut" style="height: 0.58056em; vertical-align: -0.15em;"></span><span class="mord"><span class="mord mathdefault" style="margin-right: 0.02691em;">w</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height: 0.311664em;"><span class="" style="top: -2.55em; margin-left: -0.02691em; margin-right: 0.05em;"><span class="pstrut" style="height: 2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathdefault mtight">i</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height: 0.15em;"><span class=""></span></span></span></span></span></span></span></span></span></span></font></p> 


nn.Parameter的使用：可学习权重设置

 
 

 
 
  • 1

在这里插入图片描述

更新记录

  • 2022年04月14日17:37:53

最近看有不少同学关注此博客，所以，我就找一个可以直接运行的手写数字识别代码，把可学习参数放进去，在训练时输出; 为了让大家能够对可学习参数的变化，有更好的理解。

手写数字识别代码

"""
Author: yida
Time is: 2022/3/6 09:30
this Code: 代码原文: https://www.cnblogs.com/wj-1314/p/9842719.html
- 代码: 手写数字识别, 源码参考上面的链接, 仅仅包含两个卷积层的手写数字识别 对每个卷积层设置一个权重系数w
- 可直接运行, torchvision会自动下载手写数字识别的数据集 存放在当前文件夹 ./mnist 模型保存为./model.pth
- 未实现测试功能 大家可以自行添加
- 为了便于大家更好的理解可学习参数
- 直接放到代码里面, 边训练边输出, 方便各位理解
"""
import os

import torch
import torch.nn as nn
import torchvision.datasets as normal_datasets
import torchvision.transforms as transforms
from torch.autograd import Variable

os.environ[“KMP_DUPLICATE_LIB_OK”] = “TRUE”

# 两层卷积
class CNN(nn.Module):
def init(self):
super(CNN, self).init()
# 使用序列工具快速构建
self.conv1 = nn.Sequential(
nn.Conv2d(1, 16, kernel_size=5, padding=2),
nn.BatchNorm2d(16),
nn.ReLU(),
nn.MaxPool2d(2))

    self.conv2 <span class="token operator">=</span> nn.Sequential<span class="token punctuation">(</span>
        nn.Conv2d<span class="token punctuation">(</span><span class="token number">16</span>, <span class="token number">32</span>, <span class="token assign-left variable">kernel_size</span><span class="token operator">=</span><span class="token number">5</span>, <span class="token assign-left variable">padding</span><span class="token operator">=</span><span class="token number">2</span><span class="token punctuation">)</span>,
        nn.BatchNorm2d<span class="token punctuation">(</span><span class="token number">32</span><span class="token punctuation">)</span>,
        nn.ReLU<span class="token punctuation">(</span><span class="token punctuation">)</span>,
        nn.MaxPool2d<span class="token punctuation">(</span><span class="token number">2</span><span class="token punctuation">))</span>

    self.fc <span class="token operator">=</span> nn.Linear<span class="token punctuation">(</span><span class="token number">7</span> * <span class="token number">7</span> * <span class="token number">32</span>, <span class="token number">10</span><span class="token punctuation">)</span>
    self.w <span class="token operator">=</span> nn.Parameter<span class="token punctuation">(</span>torch.ones<span class="token punctuation">(</span><span class="token number">2</span><span class="token punctuation">))</span>  <span class="token comment"># 初始化权重, 对2个卷积分别加一个权重</span>

def forward<span class="token punctuation">(</span>self, x<span class="token punctuation">)</span>:
    <span class="token comment"># 归一化权重</span>
    w1 <span class="token operator">=</span> torch.exp<span class="token punctuation">(</span>self.w<span class="token punctuation">[</span><span class="token number">0</span><span class="token punctuation">]</span><span class="token punctuation">)</span> / torch.sum<span class="token punctuation">(</span>torch.exp<span class="token punctuation">(</span>self.w<span class="token punctuation">))</span>
    w2 <span class="token operator">=</span> torch.exp<span class="token punctuation">(</span>self.w<span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">)</span> / torch.sum<span class="token punctuation">(</span>torch.exp<span class="token punctuation">(</span>self.w<span class="token punctuation">))</span>
    out <span class="token operator">=</span> self.conv1<span class="token punctuation">(</span>x<span class="token punctuation">)</span> * w1
    out <span class="token operator">=</span> self.conv2<span class="token punctuation">(</span>out<span class="token punctuation">)</span> * w2
    out <span class="token operator">=</span> out.view<span class="token punctuation">(</span>out.size<span class="token punctuation">(</span><span class="token number">0</span><span class="token punctuation">)</span>, -1<span class="token punctuation">)</span>  <span class="token comment"># reshape</span>
    out <span class="token operator">=</span> self.fc<span class="token punctuation">(</span>out<span class="token punctuation">)</span>
    <span class="token builtin class-name">return</span> out

# 将数据处理成Variable, 如果有GPU, 可以转成cuda形式
def get_variable(x):
x = Variable(x)
return x.cuda() if torch.cuda.is_available() else x

if name == ‘main’:

num_epochs <span class="token operator">=</span> <span class="token number">5</span>
batch_size <span class="token operator">=</span> <span class="token number">100</span>
learning_rate <span class="token operator">=</span> <span class="token number">0.001</span>

<span class="token comment"># 从torchvision.datasets中加载一些常用数据集</span>
train_dataset <span class="token operator">=</span> normal_datasets.MNIST<span class="token punctuation">(</span>
    <span class="token assign-left variable">root</span><span class="token operator">=</span><span class="token string">'./mnist/'</span>,  <span class="token comment"># 数据集保存路径</span>
    <span class="token assign-left variable">train</span><span class="token operator">=</span>True,  <span class="token comment"># 是否作为训练集</span>
    <span class="token assign-left variable">transform</span><span class="token operator">=</span>transforms.ToTensor<span class="token punctuation">(</span><span class="token punctuation">)</span>,  <span class="token comment"># 数据如何处理, 可以自己自定义</span>
    <span class="token assign-left variable">download</span><span class="token operator">=</span>True<span class="token punctuation">)</span>  <span class="token comment"># 路径下没有的话, 可以下载</span>

<span class="token comment"># 见数据加载器和batch</span>
test_dataset <span class="token operator">=</span> normal_datasets.MNIST<span class="token punctuation">(</span>root<span class="token operator">=</span><span class="token string">'./mnist/'</span>,
                                     <span class="token assign-left variable">train</span><span class="token operator">=</span>False,
                                     <span class="token assign-left variable">transform</span><span class="token operator">=</span>transforms.ToTensor<span class="token punctuation">(</span><span class="token punctuation">))</span>

train_loader <span class="token operator">=</span> torch.utils.data.DataLoader<span class="token punctuation">(</span>dataset<span class="token operator">=</span>train_dataset,
                                           <span class="token assign-left variable">batch_size</span><span class="token operator">=</span>batch_size,
                                           <span class="token assign-left variable">shuffle</span><span class="token operator">=</span>True<span class="token punctuation">)</span>

test_loader <span class="token operator">=</span> torch.utils.data.DataLoader<span class="token punctuation">(</span>dataset<span class="token operator">=</span>test_dataset,
                                          <span class="token assign-left variable">batch_size</span><span class="token operator">=</span>batch_size,
                                          <span class="token assign-left variable">shuffle</span><span class="token operator">=</span>False<span class="token punctuation">)</span>

model <span class="token operator">=</span> CNN<span class="token punctuation">(</span><span class="token punctuation">)</span>
<span class="token keyword">if</span> torch.cuda.is_available<span class="token punctuation">(</span><span class="token punctuation">)</span>:
    model <span class="token operator">=</span> model.cuda<span class="token punctuation">(</span><span class="token punctuation">)</span>

<span class="token comment"># 选择损失函数和优化方法</span>
loss_func <span class="token operator">=</span> nn.CrossEntropyLoss<span class="token punctuation">(</span><span class="token punctuation">)</span>
optimizer <span class="token operator">=</span> torch.optim.Adam<span class="token punctuation">(</span>model.parameters<span class="token punctuation">(</span><span class="token punctuation">)</span>, <span class="token assign-left variable">lr</span><span class="token operator">=</span>learning_rate<span class="token punctuation">)</span>

<span class="token keyword">for</span> <span class="token for-or-select variable">epoch</span> <span class="token keyword">in</span> range<span class="token punctuation">(</span>num_epochs<span class="token punctuation">)</span>:
    <span class="token keyword">for</span> i, <span class="token punctuation">(</span>images, labels<span class="token punctuation">)</span> <span class="token keyword">in</span> enumerate<span class="token punctuation">(</span>train_loader<span class="token punctuation">)</span>:
        images <span class="token operator">=</span> get_variable<span class="token punctuation">(</span>images<span class="token punctuation">)</span>
        labels <span class="token operator">=</span> get_variable<span class="token punctuation">(</span>labels<span class="token punctuation">)</span>
        outputs <span class="token operator">=</span> model<span class="token punctuation">(</span>images<span class="token punctuation">)</span>
        loss <span class="token operator">=</span> loss_func<span class="token punctuation">(</span>outputs, labels<span class="token punctuation">)</span>
        optimizer.zero_grad<span class="token punctuation">(</span><span class="token punctuation">)</span>
        loss.backward<span class="token punctuation">(</span><span class="token punctuation">)</span>
        optimizer.step<span class="token punctuation">(</span><span class="token punctuation">)</span>
        <span class="token keyword">if</span> <span class="token punctuation">(</span>i + <span class="token number">1</span><span class="token punctuation">)</span> % <span class="token number">100</span> <span class="token operator">==</span> <span class="token number">0</span>:
            print<span class="token punctuation">(</span><span class="token string">'Epoch [%d/%d], Iter [%d/%d] Loss: %.4f'</span>
                  % <span class="token punctuation">(</span>epoch + <span class="token number">1</span>, num_epochs, i + <span class="token number">1</span>, len<span class="token punctuation">(</span>train_dataset<span class="token punctuation">)</span> // batch_size, loss.item<span class="token punctuation">(</span><span class="token punctuation">))</span><span class="token punctuation">)</span>

            <span class="token comment"># 动态输出w权重变换</span>
            <span class="token keyword">for</span> name, p <span class="token keyword">in</span> model.named_parameters<span class="token punctuation">(</span><span class="token punctuation">)</span>:
                <span class="token keyword">if</span> name <span class="token operator">==</span> <span class="token string">'w'</span><span class="token builtin class-name">:</span>
                    print<span class="token punctuation">(</span><span class="token string">"特征权重: "</span>, name<span class="token punctuation">)</span>
                    w0 <span class="token operator">=</span> <span class="token punctuation">(</span>torch.exp<span class="token punctuation">(</span>p<span class="token punctuation">[</span><span class="token number">0</span><span class="token punctuation">]</span><span class="token punctuation">)</span> / torch.sum<span class="token punctuation">(</span>torch.exp<span class="token punctuation">(</span>p<span class="token punctuation">))</span><span class="token punctuation">)</span>.item<span class="token punctuation">(</span><span class="token punctuation">)</span>
                    w1 <span class="token operator">=</span> <span class="token punctuation">(</span>torch.exp<span class="token punctuation">(</span>p<span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">)</span> / torch.sum<span class="token punctuation">(</span>torch.exp<span class="token punctuation">(</span>p<span class="token punctuation">))</span><span class="token punctuation">)</span>.item<span class="token punctuation">(</span><span class="token punctuation">)</span>
                    print<span class="token punctuation">(</span><span class="token string">"w0={} w1={}"</span>.format<span class="token punctuation">(</span>w0, w1<span class="token punctuation">))</span>
                    print<span class="token punctuation">(</span><span class="token string">""</span><span class="token punctuation">)</span>

<span class="token comment"># Save the Trained Model</span>
print<span class="token punctuation">(</span><span class="token string">"训练完成..."</span><span class="token punctuation">)</span>
torch.save<span class="token punctuation">(</span>model.state_dict<span class="token punctuation">(</span><span class="token punctuation">)</span>, <span class="token string">'./model.pth'</span><span class="token punctuation">)</span>
  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40
  • 41
  • 42
  • 43
  • 44
  • 45
  • 46
  • 47
  • 48
  • 49
  • 50
  • 51
  • 52
  • 53
  • 54
  • 55
  • 56
  • 57
  • 58
  • 59
  • 60
  • 61
  • 62
  • 63
  • 64
  • 65
  • 66
  • 67
  • 68
  • 69
  • 70
  • 71
  • 72
  • 73
  • 74
  • 75
  • 76
  • 77
  • 78
  • 79
  • 80
  • 81
  • 82
  • 83
  • 84
  • 85
  • 86
  • 87
  • 88
  • 89
  • 90
  • 91
  • 92
  • 93
  • 94
  • 95
  • 96
  • 97
  • 98
  • 99
  • 100
  • 101
  • 102
  • 103
  • 104
  • 105
  • 106
  • 107
  • 108
  • 109
  • 110
  • 111
  • 112
  • 113
  • 114
  • 115
  • 116
  • 117
  • 118

【推荐阅读】

Pytorch-GPU安装教程大合集（Perfect完美系列）

bao文
关注
  • 1
    点赞
  • 踩
  • 13
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
专栏目录
基于通道权重融合特征的目标跟踪算法
05-06
目标跟踪是机器视觉领域的一个研究热点, 如何提高复杂场景下的跟踪水平是一个挑战性的问题. 以往的研究表明, 如何有效使用特征是实现跟踪的关键. 因此, 提出一种基于通道融合特征的目标跟踪算法. 该方法基于多通道相关滤波框架, 引入特征通道权重, 根据通道对响应值的贡献度调整权重, 从而构建实时特征组合. 该算法能够快速捕捉目标状态变化, 有效跟踪目标. 为了验证算法跟踪的有效性, 我们在公开数据集OTB-2015上测试算法性能, 并与多种跟踪算法进行比较. 实验结果显示, 该算法在跟踪精度、成功率上都取得较好的结果, 整体性能优于对比算法.
Dassl.pytorch:用于域自适应和半监督学习的PyTorch工具箱
03-20
达斯尔Dassl是一个工具箱，旨在研究领域适应和半监督学习（因此而命名为Dassl ）。它具有模块化设计和统一的界面，可以快速进行原型设计和新DA / SSL方法的试验。使用Dassl，只需几行代码即可实现一种新方法。您可以...
即插即用的涨点模块之特征融合（DFM）详解及代码，可应用于检测、分割、分类等各种算法领域
qq_51511878的博客
04-19 2016
来源：ISPRS2022Dense fusion module（DFM）由作者为孪生网络的变化检测所提出。双时态特征融合是变化检测的一个关键部分。这一任务难以处理有两个原因：（1）输入到孪生网络中的双时相图像在空间位置和颜色上往往存在偏差。（2）背景对象复杂多变。传统方法使用直接减法或连接来融合特征。不幸的是，尽管孪生网络通过双时像提取特征，但双时像特征之间仍存在很多不对齐的问题。也有许多研究者尝试用注意力机制来解决这个问题。然而，大多数现有的基于注意力的特征模块都会引入大量计算并消耗相当多的内存。
最新Adaptive特征融合策略，涨点又高效，想发表论文可以参考
最新发布
学姐带你玩AI的博客
05-24 1008
是一种非常高效的数据处理方法，它比传统的特征更能适应不同的数据和任务需求，也因此拥有广泛的应用前景，这种方法通过动态选择和整合来自不同层次或尺度的特征信息，不仅显著提升了模型性能，实现了的效果，还优化了特征的使用效率，帮助我们加快实验迭代的速度。尤其是ASFF方法，ASFF通过学习每个空间位置上不同层级特征的重要程度，自适应地过滤掉携带矛盾信息的特征，从而解决了特征不一致性问题。为帮助各位理解，我这次分享了，包含2024最新，开源代码已附，论文可参考创新点做了简单分析，希望可以给各位带来一些灵感。
Pytorch中nn.Parameter（）参数的使用
热门推荐
qq_48479056的博客
03-29 1万+
这两天在学习怎么对网络的两个输出f1，f2，分别得到两个权重值w1，w2，然后做一个自适应的加权融合，得到最终的结果Q。在网上百度，发现了这个函数的使用nn.Parameter（），从而记录一下。 1、含义解读： 讲的很清楚，引用！ Parameter：参数。在做神经网络的训练时，其实就是训练一个模型，这个模型就是去学习一个函数，这个函数可以准确的学习到我们想要到的东西，比如正确的对物体进行分类。函数的输入就是模型的输入：一张图像，得到的模型输出就是一个预测值。我们平时的学习率，batch大小等都是一些常见
Pytorch学习(3) —— nn.Parameter nn.ParameterList nn.ParameterDict 源码解析
小玺大大
04-04 9693
为了更好理解Pytorch基本类的实现方法，我这里给出了关于参数方面的3个类的源码详解。 此部分可以更好的了解实现逻辑结构，有助于后续代码理解，学pytorch的话这个不是必须掌握的，看不懂也没关系。 文章目录1 Parameter 参数类源码2 ParameterList 参数列表类源码3 ParameterDict 参数字典类源码总结 1 Parameter 参数类源码 此部分参考《pyto...
nn.parameter
weixin_36670529的博客
04-30 1978
class torch.nn.parameter.Parameter[source] 作为一种模型参数的张量。参数是张量的子类，在与模块s一起使用时具有一个非常特殊的特性——当它们被指定为模块属性时，它们会自动添加到参数列表中，并且会出现，例如在parameters()迭代器中。分配一个张量就没有这样的效果。这是因为人们可能想要在模型中缓存一些临时状态，比如RNN的最后一个隐藏状态。如果没有像Parameter的类，这些临时变量也要注册。 参数： data (Tensor) – 参数张量
nn.Parameter 详解 （将 nn.Parameter 用于 nn.init.kaiming_normal 初始化）
莫觉子
04-04 3198
nn.Parameter 详解 （nn.Parameter用于 nn.init.kaiming_normal 初始化）
【torch.nn.Parameter 】参数相关的介绍和使用
zyw2002的博客
12-04 6450
是继承自的子类，其主要作用是作为中的可训练参数使用。它与的区别就是会自动被认为是的可训练参数，即加入到这个迭代器中去。具体格式如下： 其中 为待传入的 ， 默认为 True。事实上， 中提供的模块中的参数均是 类，例如： 参数构造 可以看作是一个类型转换函数，将一个不可训练的类型 转换成可以训练的类型 ，并将这个 绑定到这个 里面添加的参数会被添加到列表中，会被送入优化器中随训练一起学习更新此时调用 方法会显示参数。读者可自行体会以下两端代码： 相当于把传入的数据包装成一个参数，如果要直接访问/使
PyTorch里面的torch.nn.Parameter()详解
09-18
在PyTorch中，`torch.nn.Parameter()`是一个非常关键的类，它用于创建可学习的参数。这些参数通常是神经网络模型中的权重和偏置，它们在训练过程中会被优化算法更新以最小化损失函数。本文将深入探讨`torch.nn....
同类多传感器自适应加权估计的数据级融合算法研究.pdf
09-03
针对同类多传感器测量中含有的噪声,提出了多传感器数据自适应加权融合估计算法,该算法不要求知道传 感器测量数据的任何先验知识,依据估计的各传感器的方差的变化,及时调整参与融合的各传感器的权系数,使融 合系统的均方误差始终最小,并在理论上证明了该估计算法的线性无偏最小方差性. 仿真结果表明了本算法的有 效性,其融合结果在精度、容错性方面均优于传统的平均值估计算法.
pytorch加载自定义网络权重的实现
09-18
今天小编就为大家分享一篇pytorch加载自定义网络权重的实现，具有很好的参考价值，希望对大家有所帮助。一起跟随小编过来看看吧
nn.Parameter()
qq_40657582的博客
11-29 1399
class Model(nn.Module): def __init__(self): super(Model, self).__init__() self.embedding = nn.Embedding(5, 10) self.dense_1 = nn.Linear(10, 5, bias=True) self.dense_2 = nn.Linear(5, 1, bias=True) def
Pytorch定义参数用法nn.Parameter
豆子
01-27 1982
class LayerNorm(nn.Module): #层归一化 "Construct a layernorm module (See citation for details)." def __init__(self, features, eps=1e-6): super(LayerNorm, self).__i...
基于局部占优的自适应加权平均融合算法
CJ技术空间站
07-28 3776
  以下两幅是原始图像：下面的图像是融合之后的图像：这个效果还是不错的，采用局部占优的自适应加权平均融合算法，效果比PCA的小波变换融合算法的效果要好的多。呵呵，有研究这方面的同学可以共同交流哦。邮箱：cuijingmail@163.com
行人属性分类算法：自适应权重多任务学习
manonggou的博客
03-25 4199
论文：http://www.yugangjiang.info/publication/17MM-PersonAttribute.pdf github地址：https://github.com/qiexing/adaptive_weighted_attribute Adaptively Weighted Multi-task Deep Network for Person A!ribute Cl...
pytorch中nn.Parameter（）详解
ruichuang2的博客
07-14 1034
无
用pytorch写自适应权重特征融合代码
03-31
以下是使用PyTorch编写自适应权重特征融合的示例代码： ```python import torch import torch.nn as nn import torch.nn.functional as F class FeatureFusion(nn.Module): def __init__(self, input_channels, output_channels): super(FeatureFusion, self).__init__() self.conv1 = nn.Conv2d(input_channels, output_channels, kernel_size=3, padding=1) self.conv2 = nn.Conv2d(output_channels, output_channels, kernel_size=3, padding=1) self.conv3 = nn.Conv2d(output_channels, output_channels, kernel_size=3, padding=1) self.conv4 = nn.Conv2d(output_channels, output_channels, kernel_size=3, padding=1) self.relu = nn.ReLU(inplace=True) self.weight_net = nn.Sequential( nn.Conv2d(input_channels, output_channels, kernel_size=1), nn.BatchNorm2d(output_channels), nn.ReLU(inplace=True), nn.Conv2d(output_channels, output_channels, kernel_size=1), nn.BatchNorm2d(output_channels), nn.Sigmoid() ) def forward(self, x): w = self.weight_net(x) x1 = self.conv1(x) x1 = self.relu(x1) x1 = self.conv2(x1) x1 = self.relu(x1) x2 = self.conv3(x) x2 = self.relu(x2) x2 = self.conv4(x2) x2 = self.relu(x2) x = w * x1 + (1 - w) * x2 return x ``` 在这个示例中，我们定义了一个名为FeatureFusion的模块，它接受两个特征图作为输入，并以自适应的方式将它们融合在一起。该模块包含四个卷积层，两个用于处理第一个输入特征图，两个用于处理第二个输入特征图，并且一个权重网络，它生成一个0到1之间的权重，表示对应于两个输入特征图的每个位置使用哪个特征图的比例。最后，我们使用这个权重来将两个特征图融合在一起。

“相关推荐”对你有帮助么？

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值