This contribution presents simulation results for the EE2 test 2.10 on adaptive HoG for DIMD. The method proposes modifications to the computation of the Histogram of Gradient (HoG) for DIMD. Rather than using a fixed DIMD template, an adaptive number of samples are considered to compute the HoG depending on the histogram amplitudes. An extended DIMD template is used comprising reconstructed samples further away from the current block. Also, the amplitudes are scaled with a factor which depends on the distance from the current block and on the computed directions. Finally, the number of DIMD predictors is extended for larger blocks.