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.

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

Non-Linearized Position Embedding可以展开介绍一下吗 #69

Open
bojone opened this issue Oct 13, 2023 · 3 comments
Open

Non-Linearized Position Embedding可以展开介绍一下吗 #69

bojone opened this issue Oct 13, 2023 · 3 comments

Comments

@bojone
Copy link

bojone commented Oct 13, 2023

“为此,智源团队创新提出NLPE(Non-Linearized Position Embedding,非线性位置编码)方法,在 RoPE 方法的基础上,通过调整相对位置编码、约束最大相对长度来提升模型外延能力。”

来自 https://mp.weixin.qq.com/s/ZQF4Y-kJaPKn5q69WoxmzQ 的介绍,对NLPE部分比较感兴趣。我看hf上的代码也好像没发现相关内容。

@bojone bojone changed the title Non-Linearized Position Embedding可以展开谈谈吗 Non-Linearized Position Embedding可以展开介绍一下吗 Oct 13, 2023
@isuco
Copy link
Collaborator

isuco commented Oct 14, 2023

“为此,智源团队创新提出NLPE(Non-Linearized Position Embedding,非线性位置编码)方法,在 RoPE 方法的基础上,通过调整相对位置编码、约束最大相对长度来提升模型外延能力。”

来自 https://mp.weixin.qq.com/s/ZQF4Y-kJaPKn5q69WoxmzQ 的介绍,对NLPE部分比较感兴趣。我看hf上的代码也好像没发现相关内容。

感谢关注~我们已经在准备开源代码了,预计下周会加到仓库。同时之后也会有详细的技术报告来解释NLPE的工作。

@yuanjypku
Copy link
Collaborator

感谢回复~NLPE的具体原理是基于修正attention分布的frequency-aware & position aware 位置编码修改,具体细节我们后续会发布在技术报告里

@bojone
Copy link
Author

bojone commented Dec 9, 2023

打扰一下,请问这个还有后续介绍吗?我有没有错过啥?

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

No branches or pull requests

3 participants