NVIDIA-NeMo / NVIDIA-NeMo/Automodel
TE RMS Norm numerical Instability
Open
@ZhiyuLi-Nvidia is already working on this.
Since Mar 3, 2026.
bug
- Dominant language
- Python
- Stars
- 963
- Forks
- 318
- Avg merge
- 3d 20h
- Merged PRs (30d)
- 143
Description
Describe the bug
Found TE RMS Norm is numerical instable in experiments. Discover the right version/setup for better accuracy.
Steps/Code to reproduce bug
Examples:
- Qwen2.5-7B: https://wandb.ai/nvidia/automodel-dev-zhiyul?nw=ouoluvgy6a
- llama3.2-3B: https://wandb.ai/nvidia/automodel-dev-zhiyul/workspace?nw=7nq9jmggxxm
The above 2 exps is valid and convergent only with TE RMS Norm disable.
A helpful guide on on how to craft a minimal bug report http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports.
Expected behavior
Match convergence with TE RMS Norm enabled.
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.