-
-
Notifications
You must be signed in to change notification settings - Fork 5.3k
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.
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
[V1][Bugfix] Always set enable_chunked_prefill = True for V1 #11061
Conversation
Signed-off-by: Woosuk Kwon <[email protected]>
👋 Hi! Thank you for contributing to the vLLM project. Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging. To run CI, PR reviewers can do one of these:
🚀 |
Signed-off-by: Woosuk Kwon <[email protected]>
Signed-off-by: Woosuk Kwon <[email protected]>
…oject#11061) Signed-off-by: Woosuk Kwon <[email protected]> Signed-off-by: Akshat Tripathi <[email protected]>
…oject#11061) Signed-off-by: Woosuk Kwon <[email protected]>
…oject#11061) Signed-off-by: Woosuk Kwon <[email protected]>
This PR fixes a bug in initializing the OpenAI server with V1: Currently, the server fails for some models with the error message:
ValueError: max_num_batched_tokens (2048) is smaller than max_model_len (8192).
This is becauseenable_chunked_prefill
is not properly set.