增加隐层不同层学习率设置

pom.xml

@@ -16,6 +16,15 @@
         <maven.compiler.source>1.8</maven.compiler.source>
         <maven.compiler.target>1.8</maven.compiler.target>
     </properties>
+    <dependencies>
+        <dependency>
+            <groupId>com.alibaba</groupId>
+            <artifactId>fastjson</artifactId>
+            <version>1.2.51</version>
+            <scope>test</scope>
+        </dependency>
+
+    </dependencies>
     <build>
         <pluginManagement>
             <plugins>

src/main/java/org/wlld/Test.java

@@ -1,15 +0,0 @@
-package org.wlld;
-
-import org.wlld.i.OutBack;
-
-/**
- * @author lidapeng
- * @description
- * @date 1:19 下午 2019/12/24
- */
-public class Test implements OutBack {
-    @Override
-    public void getBack(double out, int id, long eventId) {
-        System.out.println("out==" + out + ",id==" + id + ",eventId==" + eventId);
-    }
-}

src/main/java/org/wlld/imageRecognition/TempleConfig.java

@@ -116,6 +116,16 @@ public class TempleConfig {
         nerveManager.setStudyPoint(studyPoint);
     }
 
+    public void setStudyList(List<Double> list) {//设置每一层不同的学习率
+        if (studyPattern == StudyPattern.Accuracy_Pattern) {
+            //给卷积层设置层学习率
+            convolutionNerveManager.setStudyList(list);
+        } else if (studyPattern == StudyPattern.Speed_Pattern) {
+            //给全连接层设置学习率
+            nerveManager.setStudyList(list);
+        }
+    }
+
     public int getRow() {
         return row;
     }
@@ -131,6 +141,9 @@ public class TempleConfig {
     //注入模型参数
     public void insertModel(ModelParameter modelParameter) throws Exception {
         nerveManager.insertModelParameter(modelParameter);
+        if (studyPattern == StudyPattern.Accuracy_Pattern) {
+            convolutionNerveManager.insertModelParameter(modelParameter);
+        }
     }
 
     public void setCutThreshold(double cutThreshold) {

src/main/java/org/wlld/nerveCenter/NerveManager.java

@@ -31,6 +31,15 @@ public class NerveManager {
     private ActiveFunction activeFunction;
     private Map<Integer, Matrix> matrixMap = new HashMap<>();//主键与期望矩阵的映射
     private boolean isDynamic;//是否是动态神经网络
+    private List<Double> studyList = new ArrayList<>();
+
+    public List<Double> getStudyList() {//查看每一次的学习率
+        return studyList;
+    }
+
+    public void setStudyList(List<Double> studyList) {//设置每一层的学习率
+        this.studyList = studyList;
+    }
 
     public void setMatrixMap(Map<Integer, Matrix> matrixMap) {
         this.matrixMap = matrixMap;
@@ -42,7 +51,7 @@ public class NerveManager {
 
     public void setStudyPoint(double studyPoint) throws Exception {
         //设置学习率
-        if (studyPoint < 1 && studyPoint > 0) {
+        if (studyPoint <= 1 && studyPoint > 0) {
             this.studyPoint = studyPoint;
         } else {
             throw new Exception("studyPoint Values range from 0 to 1");
@@ -119,9 +128,10 @@ public class NerveManager {
     }
 
     public void insertModelParameter(ModelParameter modelParameter) throws Exception {
-        insertBpModelParameter(modelParameter);//全连接层注入参数
         if (isDynamic) {
-            insertConvolutionModelParameter(modelParameter);
+            insertConvolutionModelParameter(modelParameter);//动态神经元注入
+        } else {
+            insertBpModelParameter(modelParameter);//全连接层注入参数
         }
     }
 
@@ -278,6 +288,13 @@ public class NerveManager {
     private void initDepthNerve(boolean isMatrix) throws Exception {//初始化隐层神经元1
         for (int i = 0; i < hiddenDepth; i++) {//遍历深度
             List<Nerve> hiddenNerveList = new ArrayList<>();
+            double studyPoint = this.studyPoint;
+            if (studyList.contains(i)) {//加载每一层的学习率
+                studyPoint = studyList.get(i);
+            }
+            if (studyPoint <= 0 || studyPoint > 1) {
+                throw new Exception("studyPoint Values range from 0 to 1");
+            }
             for (int j = 1; j < hiddenNerverNub + 1; j++) {//遍历同级
                 int upNub = 0;
                 int downNub = 0;

src/main/java/org/wlld/test/Test.java

@@ -1,57 +0,0 @@
-package org.wlld.test;
-
-import org.wlld.MatrixTools.Matrix;
-import org.wlld.imageRecognition.Operation;
-import org.wlld.imageRecognition.Picture;
-import org.wlld.imageRecognition.TempleConfig;
-import org.wlld.tools.ArithUtil;
-
-/**
- * @author lidapeng
- * @description
- * @date 2:11 下午 2020/1/7
- */
-public class Test {
-    public static Matrix E;
-    public static Matrix F;
-
-    static {
-        try {
-            E = getE(true);
-            F = getE(false);
-        } catch (Exception e) {
-            e.printStackTrace();
-        }
-    }
-
-    public static void main(String[] args) throws Exception {
-        double d = ArithUtil.div(3204, 4032);
-        int a = (int) (d * 5);
-        System.out.println(a);
-    }
-
-    public static void test() throws Exception {
-
-    }
-
-    public static Matrix getE(boolean isRight) throws Exception {
-        Matrix matrix = new Matrix(5, 4);
-        String name;
-        if (isRight) {
-            name = "[10,10,10,0]#" +
-                    "[10,10,10,0]#" +
-                    "[10,10,10,0]#" +
-                    "[10,10,10,0]#" +
-                    "[10,10,10,0]#";
-        } else {
-            name = "[1,1,1,0]#" +
-                    "[1,1,1,0]#" +
-                    "[1,1,1,0]#" +
-                    "[1,1,1,0]#" +
-                    "[1,1,1,0]#";
-        }
-
-        matrix.setAll(name);
-        return matrix;
-    }
-}

src/test/java/org/wlld/HelloWorld.java

@@ -1,6 +1,7 @@
 package org.wlld;
 
-
+import com.alibaba.fastjson.JSON;
+import com.alibaba.fastjson.JSONObject;
 import org.wlld.MatrixTools.Matrix;
 import org.wlld.config.StudyPattern;
 import org.wlld.imageRecognition.Operation;
@@ -13,14 +14,32 @@ import java.util.HashMap;
 import java.util.Map;
 
 /**
- * 测试入口类!
+ * @author lidapeng
+ * @description 测试入口类
+ * @date 11:35 上午 2020/1/18
  */
 public class HelloWorld {
     public static void main(String[] args) throws Exception {
-        testPic2();
+        //testPic2();
+        testModel();
+    }
+
+    public static void testModel() throws Exception {
+        // 模型参数获取及注入
+        TempleConfig templeConfig = getTemple(true, StudyPattern.Accuracy_Pattern);
+        ModelParameter modelParameter1 = templeConfig.getModel();
+        String model = JSON.toJSONString(modelParameter1);
+        System.out.println(model);
+        TempleConfig templeConfig2 = getTemple(false, StudyPattern.Accuracy_Pattern);
+        ModelParameter modelParameter3 = JSONObject.parseObject(model, ModelParameter.class);
+        templeConfig2.insertModel(modelParameter3);
+        ModelParameter modelParameter2 = templeConfig2.getModel();
+        String model2 = JSON.toJSONString(modelParameter2);
+        System.out.println(model2);
+
     }
 
-    public static void testPic2() throws Exception {
+    public static void testPic2() throws Exception {//测试Accuracy_Pattern 模式学习
         Picture picture = new Picture();
         TempleConfig templeConfig = getTemple(true, StudyPattern.Accuracy_Pattern);
         Operation operation = new Operation(templeConfig);
@@ -53,7 +72,7 @@ public class HelloWorld {
         operation.look(wrong, 3);
     }
 
-    public static void testPic() throws Exception {
+    public static void testPic() throws Exception {//测试SPEED模式学习
         //初始化图像转矩阵类
         Picture picture = new Picture();
         //初始化配置模板类
@@ -79,13 +98,13 @@ public class HelloWorld {
         //获取模型MODLE
         ModelParameter modelParameter = templeConfig.getModel();
         //将模型MODEL转化成JSON 字符串
-        //String model = JSON.toJSONString(modelParameter);
+        String model = JSON.toJSONString(modelParameter);
         //将JSON字符串转化为模型MODEL
-        //ModelParameter modelParameter1 = JSONObject.parseObject(model, ModelParameter.class);
+        ModelParameter modelParameter1 = JSONObject.parseObject(model, ModelParameter.class);
         //初始化模型配置
         TempleConfig templeConfig1 = getTemple(false, StudyPattern.Speed_Pattern);
         //注入之前学习结果的模型MODEL到配置模版里面
-        templeConfig1.insertModel(modelParameter);
+        templeConfig1.insertModel(modelParameter1);
         //将配置模板配置到运算类
         Operation operation1 = new Operation(templeConfig1);
         //获取本地图片字节码转化成降纬后的灰度矩阵