Skip to content
This repository has been archived by the owner on Aug 30, 2024. It is now read-only.

Yarn feature #97

Merged
merged 3 commits into from
Feb 22, 2024
Merged

Yarn feature #97

merged 3 commits into from
Feb 22, 2024

Conversation

xiguiw
Copy link
Contributor

@xiguiw xiguiw commented Jan 28, 2024 •
edited
Loading

Type of Change

  1. Add YaRN rope scaling data structure and conversion read/write of the yarn parameters

How has this PR been tested?

convert data model and read/quant the mode correctly.

Dependency Change?

No.

airMeng
airMeng suggested changes Jan 29, 2024
View reviewed changes
neural_speed/models/llama/llama_yarn.cpp Outdated Show resolved Hide resolved
@xiguiw xiguiw force-pushed the yarn-feature branch 2 times, most recently from 45ee371 to 2df7c38 Compare January 31, 2024 02:25
@xiguiw xiguiw force-pushed the yarn-feature branch 2 times, most recently from 5a90dc4 to a57eb4f Compare February 7, 2024 07:37
@airMeng airMeng merged commit 8c846d6 into intel:main Feb 22, 2024
10 checks passed
@Zhenzhong1 Zhenzhong1 mentioned this pull request Feb 23, 2024
Merged
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@zhenwei-intel zhenwei-intel zhenwei-intel approved these changes

@Zhenzhong1 Zhenzhong1 Zhenzhong1 approved these changes

@intellinjun intellinjun intellinjun approved these changes

@airMeng airMeng airMeng 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.
5 participants
@xiguiw @airMeng @intellinjun @Zhenzhong1 @zhenwei-intel