NVIDIA-NeMo / NVIDIA-NeMo/RL

More helpful error message if prepare_for_*() not called

Open
#1,141 0 comments 0 reactions 0 assignees View on GitHub
good first issue UX
Dominant language
Python
Stars
2k
Forks
561
Avg merge
4d 5h
Merged PRs (30d)
145

Description

Internal report:

If someone is making large changes or writing their own training loop. Users may encounter "illegal memory access" if the model is offloaded and not moved to the GPU before that API call. We should either document or give more helpful error messages when the appropriate prepare step is not called.

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.