Skip to content

Conversation

@sfc-gh-zhwang
Copy link
Collaborator

@sfc-gh-zhwang sfc-gh-zhwang commented Oct 10, 2023

Add an environment var BERT_FREE_BUFFER_AFTER_FORWARD, if BERT_FREE_BUFFER_AFTER_FORWARD=ON, then bert model will release activation gpu memory after inference.

cc @sfc-gh-ybsat @neevaco/corvo

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants