Inversion of large-scale gravity data set is generally a challenging problem due to memory requirements and computational costs. In this study, based on VNet, we present an efficient strategy for the large-scale gravity inverse problem by simultaneously tackling several base-scale gravity data. We first construct a large number of base-scale geological models including gravity sources, with different shapes and dimensions, and also their forward model data sets. Then, the idea of semantic segmentation is used to train an inversion network. In the next step, a finite number of base-scale and similar size area of gravity data, clipped from the original large data set with a fixed stride, are fed into the trained network. Finally, the individual recovered models are combined to provide the inversion result for the whole subsurface area. The feasibility and effectiveness of the presented inversion algorithm are tested on a large-scale complicated synthetic model. The algorithm is, then, verified for the inversion of the gravity data set obtained over the Morro do Engenho complex in central Brazil. [ABSTRACT FROM AUTHOR]