Is there any difference between the subsequent rewards of the running training node and the prover node?