NVIDIA-NeMo / NVIDIA-NeMo/RL

gemma 27b slow on v2

Open
#1,097 1 comment 0 reactions 0 assignees View on GitHub
bug t-pytdensor
Dominant language
Python
Stars
2k
Forks
561
Avg merge
4d 5h
Merged PRs (30d)
145

Description

I am noticing that gemma 27b is significantly slower on the v2 backend. I don't observe such a large slowdown with other models.

Here are three sets of runs:

https://wandb.ai/nvidia/nemo-rl?nw=drf8mhln88

* green: v2
* blue, red: v1

Image

There is some in-run variance in perf (~8%), but the diff from v1 is much larger

Related automodel issue: https://github.com/NVIDIA-NeMo/Automodel/issues/432

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.