Skip to content
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

[FaqGen] Fix the metrics parse and statistics for benchmark #215

Merged
merged 6 commits into from
Dec 21, 2024
Merged

[FaqGen] Fix the metrics parse and statistics for benchmark #215

merged 6 commits into from
Dec 21, 2024
+27 −8

Conversation

wangkl2
Copy link
Collaborator

@wangkl2 wangkl2 commented Dec 20, 2024

Description

Fix the metrics parse and statistics for FaqGen benchmark:

  • Switch from json to data param to send client request for FaqGen benchmark to fit the FormData input_type.
  • Fix the input length statistics, to avoid negative numbers of Input Tokens and Input Tokens per Second.
  • Fix the output length statistics by implementing the extraction of all generated texts from streaming output, to avoid extremely large values of Output Tokens and Output Tokens per Second.
  • Update the default input prompt and set Top K=1 in sampling for faqgen, as the original default prompt does not provide any valid details for faq generation.

Issues

#189

Type of change

List the type of change like below. Please delete options that are not relevant.

  • Bug fix (non-breaking change which fixes an issue)

Dependencies

n/a

Tests

Benchmark on faqgenfixed.

@wangkl2
Use data param instead of json to send request for faqgen
8c4c478 
Signed-off-by: Wang, Kai Lawrence <kai.lawrence.wang@intel.com>
@wangkl2
Fix the input statistics for faqgen benchmark
fb8577d 
Signed-off-by: Wang, Kai Lawrence <kai.lawrence.wang@intel.com>
@wangkl2
Update the default prompt for faqgenfixed
7c92017 
Signed-off-by: Wang, Kai Lawrence <kai.lawrence.wang@intel.com>
@wangkl2
Implement the complete_response for the streaming output
33641c6 
Signed-off-by: Wang, Kai Lawrence <kai.lawrence.wang@intel.com>
@wangkl2
Set topK=1 for faqgenfixed
4c79f8a 
Signed-off-by: Wang, Kai Lawrence <kai.lawrence.wang@intel.com>
lkk12014402
lkk12014402 approved these changes Dec 20, 2024
View reviewed changes
Copy link
Collaborator

@lkk12014402 lkk12014402 left a comment

Choose a reason for hiding this comment

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

LGTM

@lvliang-intel lvliang-intel merged commit 5d717e8 into opea-project:main Dec 21, 2024
7 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@lvliang-intel lvliang-intel lvliang-intel approved these changes

@lkk12014402 lkk12014402 lkk12014402 approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@wangkl2 @lkk12014402 @lvliang-intel