NVIDIA-NeMo / NVIDIA-NeMo/RL

Nightly test llm_dpo_nanov3_30B3AB_1n4g_fsdp4ep4_automodel metric check failing

Open
#2,662 2 comments 0 reactions 1 assignee Claimed by @ashors1 View on GitHub
accuracy bug
Dominant language
Python
Stars
2k
Forks
561
Avg merge
4d 5h
Merged PRs (30d)
145

Description

**Describe the bug**

Nightly test for llm_dpo_nanov3_30B3AB_1n4g_fsdp4ep4_automodel is failing the metrics check by a close margin.

**Steps/Code to reproduce bug**

Failing off the main branch

```
│ FAIL │ data["train/preferenc… │ 0.6920857429504395 │ data["train/preferen… │
│ │ > 0.69314 │ │ > 0.69314 (condition │
│ │ │ │ evaluated to False)
```

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.