[Determinism] Add deterministic TE fused loss support
Open
Determinism
enhancement
- Dominant language
- Python
- Stars
- 17.9k
- Forks
- 4.5k
- Avg merge
- 4d 6h
- Merged PRs (30d)
- 271
Description
**Is your feature request related to a problem? Please describe.**
https://github.com/NVIDIA/Megatron-LM/pull/6965
No deterministic TE fused loss support.
Revert the change once we have deterministic TE fused loss support
Contributor guide
Assessment
This issue has not been assessed yet.