NVIDIA-NeMo / NVIDIA-NeMo/RL

High logprob errors with deepscaler recipe + megatron backend

Open
#964 1 comment 0 reactions 0 assignees View on GitHub
bug t-mcore
Dominant language
Python
Stars
2k
Forks
561
Avg merge
4d 5h
Merged PRs (30d)
145

Description

**Describe the bug**
Running the deepscaler recipe with Megatron yields very high logprob errors throughout training
lp_error

**Steps/Code to reproduce bug**

Find the config used to reproduce [here](https://github.com/NVIDIA-NeMo/RL/blob/3d81caf0f5df13732ff97b51366762f0ecd6e8ab/examples/configs/recipes/llm/grpo-deepscaler-1.5b-8K-megatron.yaml)

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.