FSDP saving state improvement
Open
@nastya236 is already working on this.
Since Apr 12, 2026.
cuda
distributed
enhancement
- Dominant language
- C++
- Stars
- 28.5k
- Forks
- 2.3k
- Avg merge
- 3d 8h
- Merged PRs (30d)
- 62
Description
Currently in FSDP skips intermediate reconstruction of tree. This results in the following issues:
- grouping for the parameters might be non-deterministic -> loaded state might be incorrect
- state tight to the number of groups -- so to the size of the communication, which is wrong design
While current strategy works deterministically if everything is fixed from run to run, however improvement is needed.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.