diff --git a/deepspeed/runtime/zero/stage_1_and_2.py b/deepspeed/runtime/zero/stage_1_and_2.py index 457fdd291bc6..c4163d6a850f 100755 --- a/deepspeed/runtime/zero/stage_1_and_2.py +++ b/deepspeed/runtime/zero/stage_1_and_2.py @@ -1069,7 +1069,7 @@ def average_tensor(self, tensor): process_group = self.dp_process_group # count = 0 for i, param_idx_in_group, param_id in self.params_in_ipg_bucket: - param = self.bit16_groups[group_idx][param_idx_in_group] + param = self.bit16_groups[group_idx][param_idx_in_group] process_group = self.dp_process_group