* add CLUSTER dataset * refine according to dongyu's comments --------- Co-authored-by: rudongyu <ru_dongyu@outlook.com>