Qiea/c-cnn
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

实现cnn计算功能，修复switchdata的bug

Browse Source
This commit is contained in:
Qiea
2024-11-11 12:28:33 +08:00
parent fe062deba5
commit 0af7f203c7
5 changed files with 150 additions and 111 deletions
Download Patch File Download Diff File Expand all files Collapse all files

135
cnn.c
View File

@@ -19,6 +19,8 @@ float* expand(const float* old_matrix, int old_matrix_length, int layer){
return new_matrix;
}
//model 模型名字
//input_matrix 输入图像
//input_matrix_length 输入图像的边长：102
@@ -81,6 +83,7 @@ float* convolution(Model model_w, Model model_b, const float* input_matrix, int
}
//num_kernels 卷积核的个数：32
//area 池化的面积：2*2
//input_matrix 输入图像
@@ -121,6 +124,7 @@ float* pooling(Model model_w, const float* input_matrix, u8 input_matrix_length)
}
void print_rslt(float* rslt, u8 input_matrix_length, u32 length){
int _tmp = 0;
printf("[0:0]");
@@ -130,119 +134,88 @@ void print_rslt(float* rslt, u8 input_matrix_length, u32 length){
printf("\n[%d:%d]",++_tmp,i+1);
}
}
printf("\r\n\r\n");
printf("\r\n");
}
int main(){
model_init();
model_write("all");
model_switchdata("data");
void cnn_run(){
//第一层：填充102 * 102
float* expand_matrix_1 = expand(data.array, 100, 1);
// print_rslt(expand_matrix_1, 102, (1*10*102));
float* conv_rlst_1 = convolution(conv1_weight,conv1_bias,expand_matrix_1, 102);
// print_rslt(conv_rlst_1, 100*0.01, (0.01*10*100));
float* pool_rslt_1 = pooling(conv1_weight, conv_rlst_1, 100);
// print_rslt(pool_rslt_1, 50, (1*50*50));
//第二层：填充32 * 52 * 52
float* expand_matrix_2 = expand(pool_rslt_1, 50, 32);
// print_rslt(expand_matrix_2, 52, (1*10*52));
float* conv_rlst_2 = convolution(conv2_weight,conv2_bias,expand_matrix_2, 52);
// print_rslt(conv_rlst_2, 50, (64*50*50));
float* pool_rslt_2 = pooling(conv2_weight, conv_rlst_2, 50);
// print_rslt(pool_rslt_2, 25, (1*25*25));
//第三层：填充 64 * 27 * 27
float* expand_matrix_3 = expand(pool_rslt_2, 25, 64);
// print_rslt(expand_matrix_2, 52, (1*52*52));
float* conv_rlst_3 = convolution(conv3_weight,conv3_bias,expand_matrix_3, 27);
print_rslt(conv_rlst_3, 25, (1*25*25));
float* pool_rslt_3 = pooling(conv3_weight, conv_rlst_3, 25);
// print_rslt(pool_rslt_3, 12, (1*12*12));
// 隐藏层参数地址
float *affine1_rslt = (float *) malloc(sizeof(float)*128);
memset(affine1_rslt, 0, sizeof(float)*128);
float affine1_temp; // 临时变量，用于存储全连接层的中间结果
{
// 隐藏层参数地址
float *affine1_rslt = (float *) malloc(sizeof(float)*128);
memset(affine1_rslt, 0, sizeof(float)*128);
float affine1_temp; // 临时变量，用于存储全连接层的中间结果
// 遍历128个神经元（假设隐藏层有128个神经元）
for(u8 n=0; n<128; n++)
{
affine1_temp = 0; // 每个神经元的输出初始化为0
// 进行矩阵乘法，将池化层输出展平为一维向量后，与全连接层权重进行点积
for(int i=0; i<(128*12*12); i++)
for(u8 n=0; n<128; n++)
{
affine1_temp = affine1_temp + pool_rslt_3[i] * fc1_weight.array[i+(128*12*12)*n];
affine1_temp = 0; // 每个神经元的输出初始化为0
// 进行矩阵乘法，将池化层输出展平为一维向量后，与全连接层权重进行点积
for(int i=0; i<(128*12*12); i++)
{
affine1_temp = affine1_temp + pool_rslt_3[i] * fc1_weight.array[i+(128*12*12)*n];
}
// 加上对应神经元的偏置
affine1_temp = affine1_temp + fc1_bias.array[n];
// 激活函数：ReLU（将小于0的值设为0）
if(affine1_temp > 0)
affine1_rslt[n] = affine1_temp; // 如果结果大于0，存入结果数组
else
affine1_rslt[n] = 0; // 否则存入0
}
// 加上对应神经元的偏置
affine1_temp = affine1_temp + fc1_bias.array[n];
// 激活函数：ReLU（将小于0的值设为0）
if(affine1_temp > 0)
affine1_rslt[n] = affine1_temp; // 如果结果大于0，存入结果数组
else
affine1_rslt[n] = 0; // 否则存入0
}
print_rslt(affine1_rslt,1,128);
// print_rslt(affine1_rslt,1,128);
float affine2_temp; // 临时变量，用于存储输出层的中间结果
float affine2_rslt[7]; // 存储输出层的结果（假设输出层有7个神经元）
float affine2_temp; // 临时变量，用于存储输出层的中间结果
float affine2_rslt[7]; // 存储输出层的结果（假设输出层有7个神经元）
// 比较输出层的最大值
float temp = -100; // 用于存储最大值的临时变量，初始化为一个非常小的值
int predict_num; // 用于存储预测的数字（对应最大值的索引）
float temp = -100; // 用于存储最大值的临时变量，初始化为一个非常小的值
int predict_num; // 用于存储预测的数字（对应最大值的索引）
// 遍历10个输出神经元（假设有10个类别）
for(int n=0; n<7; n++)
{
affine2_temp = 0; // 当前神经元的输出初始化为0
// 进行矩阵乘法，将隐藏层的输出与输出层权重进行点积
for(int i=0; i<128; i++)
for(int n=0; n<7; n++)
{
affine2_temp = affine2_temp + fc2_weight.array[i+128*n] * affine1_rslt[i];
affine2_temp = 0; // 当前神经元的输出初始化为0
// 进行矩阵乘法，将隐藏层的输出与输出层权重进行点积
for(int i=0; i<128; i++)
{
affine2_temp = affine2_temp + fc2_weight.array[i+128*n] * affine1_rslt[i];
}
// 加上对应神经元的偏置
affine2_temp = affine2_temp + fc2_weight.array[n];
affine2_rslt[n] = affine2_temp; // 存储输出层的结果
// 寻找最大值
if(temp <= affine2_rslt[n])
{
temp = affine2_rslt[n]; // 更新最大值
predict_num = n; // 记录最大值对应的类别索引
}
}
// 加上对应神经元的偏置
affine2_temp = affine2_temp + fc2_weight.array[n];
affine2_rslt[n] = affine2_temp; // 存储输出层的结果
print_rslt(affine2_rslt,7,7);
// 寻找最大值
if(temp <= affine2_rslt[n])
{
temp = affine2_rslt[n]; // 更新最大值
predict_num = n; // 记录最大值对应的类别索引
}
printf("Label is:%d\r\n",predict_num+1);
}
print_rslt(affine2_rslt,1,7);
printf("Label is:%d",predict_num+1);
return 0;
}

8
cnn.h
View File

@@ -6,13 +6,7 @@
void cnn_run(void);

12
cnn_model.c
View File

@@ -194,8 +194,12 @@ u8 model_write(char* model_name)
return 200;
}
sprintf(_path, "./dataset/%s.txt", _model -> name);
sprintf(_path, "./dataset/%s.txt", _model -> dname ? _model -> dname : _model -> name);
FILE *file = fopen(_path, "r");
if(file == NULL){
DEBUG_PRINTF("文件[%s]无法打开\r\n", _model -> dname ? _model -> dname : _model -> name);
return 199;
}
DEBUG_PRINTF("写入的模型参数名字是：%s\r\n", _model -> name);
if(_model -> dname)DEBUG_PRINTF("写入的Data数据集是：%s\r\n", _model -> dname);
@@ -227,9 +231,9 @@ u8 model_write(char* model_name)
if(_larr >= _model -> maxlength)break;
}
DEBUG_PRINTF("\r\n模型参数[%s]已写入到内存中! 模型长度为 %d\r\n",_model -> dname ? _model -> dname : _model -> name,_model -> realength);
return 1;
return 0;
}
return 0;
return 1;
}
@@ -273,7 +277,7 @@ u8 model_switchdata(char* data_name){
return 0;
}else{
u8 _res = model_write(data_name);
if (_res == 0) {
if (_res) {
DEBUG_PRINTF("Data数据集[%s]切换失败！！\r\n",data_name);
return 0;
}

2
debug.c
View File

@@ -1,7 +1,7 @@
#include "debug.h"
u8 _DEBUG = 1;
u8 _DEBUG = 0;

104
main.c
View File

@@ -1,20 +1,88 @@
////
//// Created by Qi on 2024/11/9.
////
//
//#include "cnn_model.h"
// Created by Qi on 2024/11/9.
//
//void main(){
// u8 res;
// model_init();
// model_write("all");
// model_switchdata("C1autosave00095_right_new_2");
//
//
//
//
//
//
// model_info("all");
// DEBUG_PRINTF("\r\nEnd结束");
//}
#include "cnn_model.h"
#include "cnn.h"
void run_dataset(){
char* modelusearr[] = {
"C1autosave00095_right_new_2",
"C1autosave00096_right_new_2",
"C1autosave00097_right_new_2",
"C1autosave00098_right_new_2",
"C1autosave00099_right_new_2",
"C1autosave00100_right_new_2",
"C1autosave00101_right_new_2",
"C1autosave00102_right_new_2",
"C1autosave00103_right_new_2",
"C1autosave00104_right_new_2",
"C1autosave00105_right_new_2",
"C1autosave00106_right_new_2",
"C1autosave00107_right_new_2",
"C1autosave00108_right_new_2",
"C1autosave00109_right_new_2",
"C1autosave00110_right_new_2",
"C1autosave00111_right_new_2",
"C1autosave00112_right_new_2",
"C1autosave00113_right_new_2",
"C1autosave00114_right_new_2",
"C1autosave00115_right_new_2",
"C1autosave00116_right_new_2",
"C1autosave00117_right_new_2",
"C1autosave00118_right_new_2",
"C1autosave00119_right_new_2",
"C1autosave00120_right_new_2",
"C1autosave00121_right_new_2",
"C1autosave00122_right_new_2",
"C1autosave00123_right_new_2",
"C1autosave00124_right_new_2",
"filtered_C1autosave00011_right_new",
"filtered_C1autosave00015_right_new",
"filtered_C1autosave00043_right_new",
"filtered_C1autosave00067_right_new",
"filtered_C1autosave00090_right_new",
"filtered_C1autosave00106_right_new",
"filtered_C1autosave00118_right_new",
"filtered_C1autosave00007_right_new",
"filtered_C1autosave00035_right_new",
"filtered_C1autosave00036_right_new",
"filtered_C1autosave00040_right_new",
"filtered_C1autosave00053_right_new",
"filtered_C1autosave00061_right_new",
"filtered_C1autosave00074_right_new",
"filtered_C1autosave00077_right_new",
"filtered_C1autosave00080_right_new",
"filtered_C1autosave00085_right_new",
"filtered_C1autosave00098_right_new",
"filtered_C1autosave00100_right_new",
"filtered_C1autosave00104_right_new",
"filtered_C1autosave00122_right_new",
"filtered_C1autosave00124_right_new",
"filtered_C1autosave00108_right_new",
"filtered_C1autosave00004_right_new",
"filtered_C1autosave00039_right_new",
"filtered_C1autosave00062_right_new",
};
for(int a=0;a<(sizeof(modelusearr) / sizeof(modelusearr[0]));a++){
SDRAM_USED();
model_switchdata(modelusearr[a]);
cnn_run();
}
printf("\r\n运行完成\r\n");
}
int main(){
model_init();
model_write("all");
run_dataset();
DEBUG_PRINTF("\r\nEnd结束");
}