diff --git a/PORTING/CNN/cnn.c b/PORTING/CNN/cnn.c
index 0b268bf..0fbabe7 100644
--- a/PORTING/CNN/cnn.c
+++ b/PORTING/CNN/cnn.c
@@ -41,14 +41,15 @@ float* expand(const float* old_matrix, int old_matrix_length, int layer){
 //c_rl 输出图像的边长：100
 //返回卷积的结果
 float* convolution(Model model_w, Model model_b, const float* input_matrix, int input_matrix_length){
+	DEBUG_PRINTF("卷积开始\r\n");
     // 初始化卷积层参数
     int im_l = input_matrix_length;
     int cr_l = input_matrix_length - 2;
     float conv_temp; // 临时变量，用于存储卷积计算的中间结果
 
     //用于合并前的数组，具有32*64*50*50(第二层)的大小
-    float* _conv_rlst = (float *) mymalloc(SRAMEX,sizeof (float) * model_w.channel * model_w.num_kernels * (cr_l * cr_l));
-    memset(_conv_rlst, 0, sizeof (float) * model_w.channel * model_w.num_kernels * (cr_l * cr_l));
+    float* _conv_rlst = (float *) mymalloc(SRAMEX,sizeof (float) * model_w.num_kernels * (cr_l * cr_l));
+    memset(_conv_rlst, 0, sizeof (float) * model_w.num_kernels * (cr_l * cr_l));
     //子图合并后的数组
     float* conv_rlst = (float *) mymalloc(SRAMEX,sizeof (float) * model_w.num_kernels * (cr_l * cr_l));
     memset(conv_rlst, 0, sizeof (float) * model_w.num_kernels * (cr_l * cr_l));
@@ -67,32 +68,36 @@ float* convolution(Model model_w, Model model_b, const float* input_matrix, int
                                     * model_w.array[((c+k*model_w.channel)*3*3) + (x*3+y)];
                         }
                     }
-                    _conv_rlst[(c*model_w.num_kernels*cr_l*cr_l) + (k*cr_l*cr_l) + (row*cr_l+col)] = conv_temp;
+                    _conv_rlst[(k*cr_l*cr_l) + (row*cr_l+col)] = conv_temp;
                 }
            }
         }
-    }
-    //合并子图
-    {
+        //合并子图
         for(int k=0; k < model_w.num_kernels; k++) {
             for (int row = 0; row < cr_l; row++) {
                 for (int col = 0; col < cr_l; col++) {
-                    conv_temp = 0; // 每个输出像素初始化为0
-                    for (int c = 0; c < model_w.channel; c++) {
-                        conv_temp += _conv_rlst[(c * model_w.num_kernels * cr_l * cr_l) + (k * cr_l * cr_l) + (row * cr_l + col)];
-                    }
-                    // 加上对应卷积核的偏置
-                    conv_temp += model_b.array[k];
-                    // 激活函数：ReLU（将小于0的值设为0）
-                    if (conv_temp > 0)
-                        conv_rlst[(k * (cr_l * cr_l)) + (row * cr_l) + (col)] = conv_temp; // 如果卷积结果大于0，存入结果数组
-                    else
-                        conv_rlst[(k * (cr_l * cr_l)) + (row * cr_l) + (col)] = 0; // 否则存入0
+                    conv_rlst[(k*cr_l*cr_l) + (row*cr_l+col)] += _conv_rlst[(k*cr_l*cr_l) + (row*cr_l+col)];
                 }
             }
         }
     }
+    for(int k=0; k < model_w.num_kernels; k++) {
+        for (int row = 0; row < cr_l; row++) {
+            for (int col = 0; col < cr_l; col++) {
+                conv_temp = conv_rlst[(k * cr_l * cr_l) + (row * cr_l + col)];
+                // 加上对应卷积核的偏置
+                conv_temp += model_b.array[k];
+                // 激活函数：ReLU（将小于0的值设为0）
+                if (conv_temp > 0)
+                    conv_rlst[(k * (cr_l * cr_l)) + (row * cr_l) + (col)] = conv_temp; // 如果卷积结果大于0，存入结果数组
+                else
+                    conv_rlst[(k * (cr_l * cr_l)) + (row * cr_l) + (col)] = 0; // 否则存入0
+            }
+        }
+    }
     myfree(SRAMEX,_conv_rlst);
+    _conv_rlst = NULL;
+	DEBUG_PRINTF("卷积结束\r\n");
     return conv_rlst;
 }
 
@@ -268,11 +273,14 @@ float* generateMatrix(Model model, const float* value)
 }
 
 void cnn_run(){
+	DEBUG_PRINTF("开始CNN\r\n");
     float value[3] = {0};
     calculate_statistics(data,&value[0]);
     if (check_threshold(data,&value[0])){
         float* _data = generateMatrix(data,&value[0]);
 
+
+		DEBUG_PRINTF("开始第一层\r\n");
         //第一层：填充102 * 102
         float* expand_matrix_1 = expand(_data, 100, 1);
         float* conv_rlst_1 = convolution(conv1_weight,conv1_bias,expand_matrix_1, 102);
@@ -285,6 +293,8 @@ void cnn_run(){
         myfree(SRAMEX,conv_rlst_1);
         conv_rlst_1 = NULL;
 
+
+		DEBUG_PRINTF("开始第二层\r\n");
         //第二层：填充32 * 52 * 52
         float* expand_matrix_2 = expand(pool_rslt_1, 50, 32);
         float* conv_rlst_2 = convolution(conv2_weight,conv2_bias,expand_matrix_2, 52);
@@ -297,6 +307,8 @@ void cnn_run(){
         myfree(SRAMEX,conv_rlst_2);
         conv_rlst_2 = NULL;
 
+
+		DEBUG_PRINTF("开始第三层\r\n");
         //第三层：填充 64 * 27 * 27
         float* expand_matrix_3 = expand(pool_rslt_2, 25, 64);
         float* conv_rlst_3 = convolution(conv3_weight,conv3_bias,expand_matrix_3, 27);
@@ -309,7 +321,10 @@ void cnn_run(){
         myfree(SRAMEX,conv_rlst_3);
         conv_rlst_3 = NULL;
 
+		DEBUG_PRINTF("开始第四层\r\n");
         float* affine1_rslt = hidden(pool_rslt_3);
+
+		DEBUG_PRINTF("开始第五层\r\n");
         float* affine2_rslt = output(affine1_rslt);
 
         myfree(SRAMEX,pool_rslt_3);
diff --git a/PORTING/CNN/cnn_model.c b/PORTING/CNN/cnn_model.c
index d749571..ce6ddb8 100644
--- a/PORTING/CNN/cnn_model.c
+++ b/PORTING/CNN/cnn_model.c
@@ -165,8 +165,8 @@ u8 model_write(char* model_name)
 			u8 _times=0;
 			u32 _larr = 0;
             u8 _len = strlen(model_name);
-			char _path[_len+1+7+30];
-            char _datapath[_len+1+7+30];
+			char _path[_len+1+7+35];
+            char _datapath[_len+1+7+35];
             char _fstr[READLENGTH+1] = {0};
 			char _sstr[2];
 			int progress;
@@ -284,7 +284,7 @@ u8 model_read(char* model_name, u32 start, u32 end, u32 gap){
 
 u8 model_switchdata(char* data_name){
 	u8 _len = strlen(data_name);
-	char _path[_len+1+7];
+	char _path[_len+1+7+35];
 	if(data.array != NULL)modelmym_free("data");
 	sprintf(_path, "0:/dataset/_data/%s.txt",data_name);
 	if(f_open(file,(const TCHAR*)_path,1)){
diff --git a/PORTING/CNN/debug.c b/PORTING/CNN/debug.c
index 00ab347..d3c1b99 100644
--- a/PORTING/CNN/debug.c
+++ b/PORTING/CNN/debug.c
@@ -2,7 +2,7 @@
 #include "led.h"
 
 
-u8 _DEBUG = 0;
+u8 _DEBUG = 1;