To solve least squares problems. QR decomposition with householder is used.
These are the templates to configure the function.
Parameters:
| rowA_num | describes the matrixA number of rows. |
| colA_num | describes the matrixA number of columns. |
| rowB_num | describes the vector dim. |
| colB_num | describes the number vectorB, more than one vetorB is supported. |
template < int rowA_num, int colA_num, int rowB_num, int colB_num > class LSTQR_Graph: public graph // fields kernel k0[colA_num-1] kernel k1 input_port din0 input_port din1 output_port dout0 output_port dout1 input_port column_id[colA_num]