不能在向量中的類物件上呼叫非靜態函式

我正在用自己的 Matrix 類構建自己的神經網路。

我試圖swishMatrix()在Matrix2D類物件上使用該函式，然后再將其添加到vector<Matrix2D>變數中。但是我收到這個錯誤，我不知道為什么。->no matching function for call to 'std::vector<Matrix2D>::push_back(int)'|

當我swishMatrix()在普通Matrix2D物件上使用它時，它作業正常。

這是Matrix2D課堂

class Matrix2D{
    public:
        int rows;
        int columns;
        vector<vector<float> > matrix;
        Matrix2D() = default;
        Matrix2D(int x, int y){
            rows = x;
            columns = y;
            for (int i = 0; i < rows; i  ) {
                vector<float> v1;
                for (int j = 0; j < columns; j  ) {
                    v1.push_back(0);
                }
                matrix.push_back(v1);
            }
        }
        swishMatrix(){            
            for (int i = 0; i < rows; i  ) {
                for (int j = 0; j < columns; j  ) {
                    matrix[i][j] = matrix[i][j] * sigmoid(matrix[i][j]);
                }
            }            
        }        
        //Here there's a lot of static functions for matrix operations
};

這是Neural Network課堂

class NeuralNewtork{
    public:
       
        //A lot more declaration here but not important
        Matrix2D first_hidden_weights;

        Matrix2D input_nodes;

        vector<Matrix2D> hidden_weights;      

        vector<Matrix2D> hidden_biases; 

        vector<Matrix2D> activated_hidden_nodes;       

        NeuralNewtork(int input_nodes, int hidden_layers, int hidden_nodes, int action_nodes){
            first_hidden_weights = Matrix2D(numberof_hidden_nodes, numberof_input_nodes);
            first_hidden_weights.randomizeMatrix();            

            hidden_weights.reserve(numberof_hidden_layers-1);
            for (int i=0; i<numberof_hidden_layers-1; i  ){
                hidden_weights.push_back(Matrix2D(numberof_hidden_nodes, numberof_hidden_nodes));
                hidden_weights.back().randomizeMatrix();
            }           

            hidden_biases.reserve(numberof_hidden_layers);
            for (int i=0; i<numberof_hidden_layers; i  ){
                hidden_biases.push_back(Matrix2D(numberof_hidden_nodes, 1));
                hidden_biases.back().randomizeMatrix();
            }     
            //There are more declerations here but they aren't important for this problem   
        }

        feedForward(Matrix2D input){
            input_nodes = input;

        
            for(int i = 0; i < numberof_hidden_layers 1; i  ){
                if(i==0){                    
                    activated_hidden_nodes.push_back(Matrix2D::matrixAddition(Matrix2D::matrixMultiplication(first_hidden_weights, input_nodes), hidden_biases[0]).swishMatrix());
                    //This is the line where I get the error
                    //no matching function for call to 'std::vector<Matrix2D>::push_back(int)'|
                    
                }
                if(i!=0 && i!=numberof_hidden_layers){
                    activated_hidden_nodes.push_back(Matrix2D::matrixAddition(Matrix2D::matrixMultiplication(hidden_weights[i-1], activated_hidden_nodes[i-1]), hidden_biases[i]).swishMatrix());  
                    //This is also a line where I get the error
                    //no matching function for call to 'std::vector<Matrix2D>::push_back(int)'|
                }
                if(i==numberof_hidden_layers){
                    //Not important
                }
            }
        }

我可能錯過了代碼的某些部分，很難保持簡短，但所有需要的變數都已正確分配。

uj5u.com熱心網友回復：

簡單的解決方法是swishMatrix再次回傳矩陣物件：

Matrix2D& swishMatrix(){            
    for (int i = 0; i < rows; i  ) {
         for (int j = 0; j < columns; j  ) {
              matrix[i][j] = matrix[i][j] * sigmoid(matrix[i][j]);
         }
    }     
    return *this;       
}  

永遠不要在方法上沒有明確回傳型別的情況下再次撰寫代碼。

標籤：C 功能 班级

上一篇：如何要求用戶在我的類和物件中輸入C 中的輸入

下一篇：有沒有更簡單的方法來從類中參考Javascript類方法？