deepspeedai / deepspeedai/DeepSpeed

[Bug] - summing over DP instead of MP

Open
#2,094 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
43.1k
Forks
5k
Avg merge
4d 15h
Merged PRs (30d)
112

Description

Hi Deepspeed team!

Should this be summing over the MP group rather than the DP group? The comment says MP and the code does DP.

https://github.com/microsoft/DeepSpeed/blob/b3388e1418e80141a902886dec39127b851a1f0a/deepspeed/runtime/zero/stage_1_and_2.py#L1156

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start at deepspeed/runtime/zero/stage_1_and_2.py around line 1156 and read the adjacent comment and group-reduction logic. Verify whether the reduction should use the model-parallel or data-parallel group, then validate the corrected behavior with the relevant ZeRO distributed tests; done means the implementation matches the documented group.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
distributed-systems
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.