[Fix] Consume much more gpt memory running eval_rm

LAION-AI / Open-Assistant

OpenAssistant is a chat-based assistant that understands tasks, can interact with third-party systems, and retrieve information dynamically to do so.

https://open-assistant.io

Apache License 2.0

36.92k stars 3.22k forks source link

[Fix] Consume much more gpt memory running eval_rm #3614

Closed SingL3 closed 12 months ago

SingL3 commented 1 year ago

Fix #3611. Still debugging or model_training.

github-actions[bot] commented 1 year ago

:x: pre-commit failed. Please run pre-commit run --all-files locally and commit the changes. Find more information in the repository's CONTRIBUTING.md