KEMBAR78
Fix config for Falcon by WoosukKwon · Pull Request #1164 · vllm-project/vllm · GitHub
Skip to content

Conversation

@WoosukKwon
Copy link
Collaborator

Currently, vLLM does not read the number of KV heads correctly and thus returns gibberish outputs for Falcon-40B model. This PR fixes the bug.

Copy link
Member

@zhuohan123 zhuohan123 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! Thanks for the fix!

@WoosukKwon WoosukKwon changed the title Fix Falcon-40b Fix config for Falcon Sep 24, 2023
@WoosukKwon WoosukKwon merged commit 9f6be86 into main Sep 24, 2023
@WoosukKwon WoosukKwon deleted the fix-falcon branch September 24, 2023 00:38
hongxiayang pushed a commit to hongxiayang/vllm that referenced this pull request Feb 13, 2024
jikunshang pushed a commit to jikunshang/vllm that referenced this pull request Apr 25, 2025
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