8000 delete prefix cache upon refit · Issue #97 · NVIDIA/NeMo-RL · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
8000

delete prefix cache upon refit #97

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
gshennvm opened this issue Mar 31, 2025 · 0 comments · Fixed by #112
Closed

delete prefix cache upon refit #97

gshennvm opened this issue Mar 31, 2025 · 0 comments · Fixed by #112
Assignees
@parthchadha

Comments

@gshennvm
Copy link
Collaborator

currently it seems vllm still loads the previous model's kv cache upon engine wakeup. it is not clear to me atm if it uses the stale kv cache upon refit. we should investigate if we need to clear it or not
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@parthchadha parthchadha

Labels
None yet
Projects
NeMo RL v0.2
Status: Done
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

fix: reset prefix cache when sleep is called to ensure prefix cache i… NVIDIA/NeMo-RL
2 participants
@parthchadha @gshennvm
0