deepseek-ai/DeepSeek-R1
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepSeek-R1.git synced 2026-11-05 00:47:42 +00:00
Code Issues 45 Projects Packages Wiki Activity

[GH-ISSUE #464] 请问本地推理deepseek r1需要多少硬件资源？ #371

New issue
Closed
opened 2026-10-26 21:20:13 +00:00 by bsakaguchi · 7 comments
bsakaguchi commented 2026-10-26 21:20:13 +00:00
Owner
Copy link

Originally created by @lsnls on GitHub (Feb 20, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-R1/issues/464

你好，deepseek是一个非常棒的工作！

请问本地推理deepseek r1需要多少显存空间？还有多少内存需求呢？

谢谢！

Originally created by @lsnls on GitHub (Feb 20, 2025). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-R1/issues/464 你好，deepseek是一个非常棒的工作！ 请问本地推理deepseek r1需要多少显存空间？还有多少内存需求呢？ 谢谢！
bsakaguchi 2026-10-26 21:20:13 +00:00
bsakaguchi commented 2026-10-26 21:20:16 +00:00
Author
Owner
Copy link

@JiwenJ commented on GitHub (Feb 20, 2025):

根据我的经验，至少要四台机器（sglang推理），我的是4台A100。内存的话，一般的机器都会配1T左右的内存。

<!-- gh-comment-id:2671221263 --> @JiwenJ commented on GitHub (Feb 20, 2025): 根据我的经验，至少要四台机器（sglang推理），我的是4台A100。内存的话，一般的机器都会配1T左右的内存。
bsakaguchi commented 2026-10-26 21:20:16 +00:00
Author
Owner
Copy link

@lcgogo commented on GitHub (Feb 21, 2025):

看多大的，7b 的一般家用电脑都行，后面越大越耗资源，671b 满血的要几百G显存。
这有个开源的部署 671b 的裁剪之后的，能在单卡 4090 跑起来
https://github.com/kvcache-ai/ktransformers

<!-- gh-comment-id:2673144045 --> @lcgogo commented on GitHub (Feb 21, 2025): 看多大的，7b 的一般家用电脑都行，后面越大越耗资源，671b 满血的要几百G显存。 这有个开源的部署 671b 的裁剪之后的，能在单卡 4090 跑起来 https://github.com/kvcache-ai/ktransformers
bsakaguchi commented 2026-10-26 21:20:16 +00:00
Author
Owner
Copy link

@lsnls commented on GitHub (Feb 21, 2025):

看多大的，7b 的一般家用电脑都行，后面越大越耗资源，671b 满血的要几百G显存。 这有个开源的部署 671b 的裁剪之后的，能在单卡 4090 跑起来 https://github.com/kvcache-ai/ktransformers

你好，ktransformer的文档提到，它对内存的要求是382GB，这个内存要求好像太高了。请问是我的理解有问题吗？
Image

<!-- gh-comment-id:2674360465 --> @lsnls commented on GitHub (Feb 21, 2025): > 看多大的，7b 的一般家用电脑都行，后面越大越耗资源，671b 满血的要几百G显存。 这有个开源的部署 671b 的裁剪之后的，能在单卡 4090 跑起来 https://github.com/kvcache-ai/ktransformers 你好，ktransformer的文档提到，它对内存的要求是382GB，这个内存要求好像太高了。请问是我的理解有问题吗？ ![Image](https://github.com/user-attachments/assets/47977a33-50c6-420f-aa7c-f614cf993cac)
bsakaguchi commented 2026-10-26 21:20:16 +00:00
Author
Owner
Copy link

@yang-xiao-test commented on GitHub (Feb 25, 2025):

请教以下 对CPU 的 要求高吗？本地部署了7B ～32B 问题回复速度超级慢慢 显存基本没在用.....

<!-- gh-comment-id:2680216105 --> @yang-xiao-test commented on GitHub (Feb 25, 2025): 请教以下 对CPU 的 要求高吗？本地部署了7B ～32B 问题回复速度超级慢慢 显存基本没在用.....
bsakaguchi commented 2026-10-26 21:20:16 +00:00
Author
Owner
Copy link

@pkghealth commented on GitHub (Mar 20, 2025):

Yang-xiao-test: if you use the app version on your phone and ask it how to setup. It will instruct you how to do so. If you don’t setup the compute function for your GPU then when you initialize it, it uses RAM and CPU instead of GPU. The CPU & RAM are much slower but can provide more capacity. You need to setup ROCm for AMD or whatever NVIDIA is using for GPU computing.
lsnls: I have heard 512 RAM or VRAM is necessary to use the full module unrestricted.

<!-- gh-comment-id:2741749815 --> @pkghealth commented on GitHub (Mar 20, 2025): Yang-xiao-test: if you use the app version on your phone and ask it how to setup. It will instruct you how to do so. If you don’t setup the compute function for your GPU then when you initialize it, it uses RAM and CPU instead of GPU. The CPU & RAM are much slower but can provide more capacity. You need to setup ROCm for AMD or whatever NVIDIA is using for GPU computing. lsnls: I have heard 512 RAM or VRAM is necessary to use the full module unrestricted.
bsakaguchi commented 2026-10-26 21:20:16 +00:00
Author
Owner
Copy link

@github-actions[bot] commented on GitHub (Apr 22, 2025):

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you believe this issue is still relevant, please leave a comment to keep it open. Thank you for your contributions!

<!-- gh-comment-id:2819747277 --> @github-actions[bot] commented on GitHub (Apr 22, 2025): This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you believe this issue is still relevant, please leave a comment to keep it open. Thank you for your contributions!
bsakaguchi commented 2026-10-26 21:20:16 +00:00
Author
Owner
Copy link

@github-actions[bot] commented on GitHub (May 27, 2025):

false

<!-- gh-comment-id:2910794993 --> @github-actions[bot] commented on GitHub (May 27, 2025): false
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
v1.0.0
Labels
Clear labels   
bug

   
closed-as-stale

   
closed-as-stale

   
enhancement

   
pull-request
Mirrored from GitHub Pull Request

  
stale

   
stale
No labels
bug
closed-as-stale
closed-as-stale
enhancement
pull-request
stale
stale
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/DeepSeek-R1#371
No description provided.