diff --git a/ding/reward_model/rnd_reward_model.py b/ding/reward_model/rnd_reward_model.py index 076a8ccb23..ca153492ae 100644 --- a/ding/reward_model/rnd_reward_model.py +++ b/ding/reward_model/rnd_reward_model.py @@ -168,6 +168,7 @@ def clear_data(self, iter: int) -> None: self.cfg, 'clear_buffer_per_iters' ), "Reward Model does not have clear_buffer_per_iters, \ if you want to clear buffer, you need to add this attribute in config." + if iter % self.cfg.clear_buffer_per_iters == 0: self.train_obs.clear()