Understanding phenomena in systems of many interacting quantum particles, known as quantum many-body systems, is one of the most sought-after objectives in contemporary physics research. The challenge of simulating such systems lies in the extensive resources required for exactly modeling quantum wave-functions, which grows exponentially with the number of particles. Recently, neural networks were demonstrated to be a promising approximation method of quantum wave functions. However, thus far, this approach was mostly focused on more traditional architectures such as Restricted Boltzmann Machines and small fully-connected networks. In this talk, we propose a method for scaling this approach to support large modern architecture. Though significantly more expressive, such architectures do not lend themselves to the conventional methods for employing neural networks for simulating quantum systems. A key part of the simulation is to sample according to the underlying distribution of particle configurations. Current methods rely on Markov-Chain Monte-Carlo sampling, which is too expensive for use with modern architectures, effectively limiting their usable size and capacity. Inspired by recent generative models, we propose a specialized deep convolutional architecture that supports efficient and exact sampling, completely circumventing the need for Markov Chain sampling. We demonstrate our approach can obtain accurate results on larger system sizes than those currently accessible to other neural-network representation of quantum states.