Skip to content

Small update to vllm batch#419

Merged
yunfeng-scale merged 1 commit intomainfrom
yunfeng-tqdm
Jan 17, 2024
Merged

Small update to vllm batch#419
yunfeng-scale merged 1 commit intomainfrom
yunfeng-tqdm

Conversation

@yunfeng-scale
Copy link
Copy Markdown
Contributor

@yunfeng-scale yunfeng-scale commented Jan 17, 2024

Pull Request Summary

  1. use vllm 0.2.5 since 0.2.7 seems slower for the workload i tested
  2. add tqdm

Test Plan and Usage Guide

How did you validate that your PR works correctly? How do you run or demo the code? Provide enough detail so a reviewer can reasonably reproduce the testing procedure. Paste example command line invocations if applicable.

@yunfeng-scale yunfeng-scale requested a review from a team January 17, 2024 20:58
Copy link
Copy Markdown
Member

@yixu34 yixu34 left a comment

Choose a reason for hiding this comment

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

Is there a vLLM github issue about this perf regression?

@yunfeng-scale
Copy link
Copy Markdown
Contributor Author

Is there a vLLM github issue about this perf regression?

vllm-project/vllm#2367
also our scenario is a corner case: max_toke_num = 1 for arc_challenge dataset

@yunfeng-scale yunfeng-scale merged commit db11cd7 into main Jan 17, 2024
@yunfeng-scale yunfeng-scale deleted the yunfeng-tqdm branch January 17, 2024 22:31
@yunfeng-scale yunfeng-scale mentioned this pull request Mar 6, 2024
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.

3 participants