Algorithm 1 Adam
Input: $s_0$—initial solution, $\eta$, ${\beta }_1$, ${\beta }_2$—learning step sizes, $f$
1.	set iteration number $t$ to 0
2.	initialize the $c_0$ and $v_0$ tensors filled with zeros
3.	set $ϵ={10}^{-8}$
4.	while the stop condition is not met do:
5.	$t=t+1$
6.	$c_t={\beta }_1\times c_{t-1}+\left(1-{\beta }_1\right)\times \nabla f\left(s_{t-1}\right)$
7.	$v_t=v_{t-1}-\left(1-{\beta }_2\right)\times \left(v_{t-1}-\nabla f^2\left(s_{t-1}\right)\right)$
8.	$s_t=\eta \times c_t\times {\left(\sqrt{v_t}+ϵ\right)}^{-1}$
9.	end loop
10.	return $s_t$