Code associated with initial submission of "Imitating Driver Behavior with Generative Adversarial Networks". Tagged files include four policies (bc_mlp, bc_gru, gail_mlp, gail_gru), splits for training and testing data (NGSIM_train_test_split.h5), and expert training trajectories used during training (core1_temp0_well1_neig0_carl1_roal0_clrr1_mtl100_clb20_rlb20_rll2_clmr100_rlmr50_seed456).