remove useless comments

microsoft · Sep 18, 2024 · b1231c4 · b1231c4
1 parent 5a44f8c
commit b1231c4
Showing 1 changed file with 0 additions and 3 deletions.
diff --git a/deepspeed/runtime/zero/stage_1_and_2.py b/deepspeed/runtime/zero/stage_1_and_2.py
@@ -1070,9 +1070,6 @@ def average_tensor(self, tensor):
 
                 process_group = self.dp_process_group
 
-                #Averages gradients at parameter level if ipg has a moe param
-                #Otherwise averaging is done at the entire buffer level at the end of the loop
-                # MoE param have different groups
                 if self.ipg_bucket_has_moe_params:
                     process_group = self.expert_dp_process_group[param.group_name] if is_moe_param(
                         param) else self.dp_process_group