Designing materials using a combinatorial approach to achieve desired mechanical properties requires multiple iterations. The cost of conducting experiments for each iteration to determine the material properties can be prohibitively high. One alternative is to use computational models, but running high-fidelity physics-based simulations can take a significant amount of time to reach a solution. A faster approach is to use deep learning (DL) based models as surrogate models to predict the target parameters. In the present work, we focus on laser-driven spall experiments on thin foils. The laser-driven spall tests can be classified as very high strain rate experiments (∼10^6). These experiments are useful in estimating the spall strength of the material. The objective of this work is to develop DL-based methods to predict the spall strength of materials and other parameters based on microstructure inputs. To achieve this, we shall use simulation results obtained from crystal plasticity finite element simulations as ground truth for our analyses.