Let $\{X_n;n\ge 1\}$ be a sequence of independent and identically distributed random variables in a regular sub-linear expectation space $(Ω,\mathscr{H},\widehat{\mathbb E})$ with the finite Choquet expectation, upper mean $\overlineμ $ and lower mean $\underlineμ $. Then for any Borel-measurable function $\varphi(x_1,\ldots,x_d)$ on $\mathbb R^d$ or continuous function $\varphi(x_1,x_2,\ldots)$ on $\mathbb R^{\mathbb N}$, $\sum_{i=1}^n X_i/n$ converges to $\underlineμ\wedge \varphi(X_1,X_2,\ldots)\wedge \overlineμ$ with upper capacity $1$. The limits of $\sum_{i=1}^nX_i/n$ can be with upper capacity 1 also a random set with boundaries being continuous functions or finite-dimensional Borel-measurable functions of $(X_1, X_2,\ldots)$.