The next generation communication systems demand a tremendous increase in the existing data rate for various applications. One solution to this issue is the use of Large Intelligent Surfaces (LIS) which consists of a panel on which reflecting elements are mounted. Their main role is to redirect the incident electromagnetic signal to the desire user. This in turn increases the strength of the received signal which increases the quality of reception, providing an enhanced Quality of Service (QoS). The Machine Learning Algorithms have been implemented for various aspects of LIS implementation like channel estimation, calculation of discrete phase shifts to name a few. Here, a Deep Learning (DL) model is proposed to evaluate the Signal to Interference Noise Ratio (SINR) for a LIS assisted communication system. The effect of Adam, SGD, RMSProp, Adamax and Adadelta optimizers on the DL model is studied. The loss function considered is Mean Squared Error (MSE). The simulation results indicate that the DL model with Adam Optimizer gives accuracy of 97.45% which is better as compared to other optimizers.