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

建议 支持 DeepSeek-R1 #2778

Open
wangyongpenga opened this issue Jan 22, 2025 · 9 comments
Open

建议 支持 DeepSeek-R1 #2778

wangyongpenga opened this issue Jan 22, 2025 · 9 comments
Labels
Milestone

Comments

@wangyongpenga
Copy link

Feature request / 功能建议

建议 支持 DeepSeek-R1

Motivation / 动机

DeepSeek-R1 蒸馏小模型

Your contribution / 您的贡献

DeepSeek-R1 https://github.com/deepseek-ai/DeepSeek-R1

@XprobeBot XprobeBot added this to the v1.x milestone Jan 22, 2025
@refeiner
Copy link

确实需要

@qinxuye
Copy link
Contributor

qinxuye commented Jan 24, 2025

Deepseek v3 和 r1 对资源的要求非常高。

我们会先支持 distill 模型。

Copy link

This issue is stale because it has been open for 7 days with no activity.

@github-actions github-actions bot added the stale label Jan 31, 2025
@qinxuye qinxuye removed the stale label Jan 31, 2025
@menkeyi001
Copy link

坐地等,有这些资源的公司也挺多

@shabooboo86
Copy link

shabooboo86 commented Feb 6, 2025

@qinxuye 建议优先支持DeepSeek-R1-GGUF。资源使用量下降不少。

Image

@gahoo
Copy link

gahoo commented Feb 6, 2025

@qinxuye 建议优先支持DeepSeek-R1-GGUF。资源使用量下降不少。

Image

https://unsloth.ai/blog/deepseekr1-dynamic

This article explains how they shrink the size, which looks promising. I think it's worth trying.

@gs80140
Copy link

gs80140 commented Feb 6, 2025

坐等啊, 需要多少资源, 列出来, 方便准备呀.

@MOON1234567890A
Copy link

我自己下载了gguf,但是不支持gpu推理吗,只有cpu在跑

@Icedcocon
Copy link

Deepseek R1 使用vllm部署时需要至少2个节点,vllm原生提供了分布式部署的能力,但是目前看xinference的架构应该没有考虑到对vllm夸节点TP和PP的支持。估计短时间内很难支持vllm全量部署DeepSeek R1 了。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

10 participants