Hi! Is there any way to run evaluation twice on two different sets? Currently I have a workaround that set “evaluation_num_episodes” as 2*N, and index the running times, if index is less than N, evaluate on the first dataset, if index is larger than N, evalutate on the second dataset, but its too slow because I have to set “num_envs_per_worker” to 1 in order to keep index in order. Is there any way to do evaluation twice?
Thank you so much!