contain the output from Amdam's Code, iteration = 300
loading data to matrix A and initialize X, Yufeng Shen
accelerated projected gradient descent with Huber cost function, He Liu
read fortran output, Yufeng Shen
outputs from Huber_AccProjGD.ipynb cap_vec and energy from different iterations from different method, for futher comaparison
zeroth output from Adam's code or just norm vector n used for calculate energy from cap_vec, e = cap_vec*n