Fullerene-MD: An in-house-developed computational material dataset for three well-known fullerene molecules, namely C60 (Ih), C70 (D5h), and C72 (D6h). This dataset consists of the optimized and 3000 deformed structures derived from molecular dynamics simulations (1000 structures for each molecule), with their corresponding HOMO and LUMO energies obtained using DFT calculations similar to those used in the QM9 dataset [1]. Pt/graphene-MD: An in-house-developed computational material dataset for a system of a platinum atom adsorbed on a graphene flake terminated by hydrogen atoms [2]. This dataset comprises approximately 21,000 optimized and deformed structures derived by molecular dynamics simulations, wherein the adsorption energies are obtained using DFT calculations. Experiments are performed on this dataset to evaluate the predictive capability of the deformation energies of the structures. [1] Sci Data 1, 140022 (2014); https://doi.org/10.1038/sdata.2014.22 [2] Phys. Rev. B 79, 115426 (2009); https://link.aps.org/doi/10.1103/PhysRevB.79.115426