THUDM / THUDM/slime

[feature] support megatron-fsdp

Open
#767 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
8.5k
Forks
1.3k
Avg merge
5h 36m
Merged PRs (30d)
22

Description

If we can enable --use-megatron-fsdp, we can save the GPUs from TP to CP and enable training longer sequences with the same number of GPUs. Currently it seems that we at least need to upgrade megatron for such feature.

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

No files, tests, or entry points are named. Start by locating where the --use-megatron-fsdp option would be defined and how the current Megatron integration handles tensor and context parallelism; review the required Megatron upgrade. Done means the option enables FSDP training and longer sequences with the same GPU count, with coverage for the new path.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.