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 #347] 请问：如何用r1模型微调或蒸馏自己的数据？还是用传统的lora? 官方好像没有介绍 #269

New issue
Closed
opened 2026-10-26 21:18:20 +00:00 by bsakaguchi · 6 comments
bsakaguchi commented 2026-10-26 21:18:20 +00:00
Owner
Copy link

Originally created by @alexhmyang on GitHub (Feb 10, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-R1/issues/347

谢谢

Originally created by @alexhmyang on GitHub (Feb 10, 2025). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-R1/issues/347 谢谢
bsakaguchi 2026-10-26 21:18:20 +00:00
bsakaguchi commented 2026-10-26 21:18:22 +00:00
Author
Owner
Copy link

@deepsicker commented on GitHub (Feb 10, 2025):

知识蒸馏是一种模型压缩技术，其核心思想是将大型教师模型的知识传递给较小的学生模型。具体而言，学生模型通过学习教师模型对输入数据的软标签（即概率分布），从而在保持较小规模的同时，尽可能接近教师模型的性能。这种方法常用于在资源受限的环境中部署高效模型

<!-- gh-comment-id:2646892418 --> @deepsicker commented on GitHub (Feb 10, 2025): 知识蒸馏是一种模型压缩技术，其核心思想是将大型教师模型的知识传递给较小的学生模型。具体而言，学生模型通过学习教师模型对输入数据的软标签（即概率分布），从而在保持较小规模的同时，尽可能接近教师模型的性能。这种方法常用于在资源受限的环境中部署高效模型
bsakaguchi commented 2026-10-26 21:18:22 +00:00
Author
Owner
Copy link

@KDD2018 commented on GitHub (Feb 11, 2025):

同问，另外，基于满血版R1微调或者蒸馏具体需要什么配置？

<!-- gh-comment-id:2650311351 --> @KDD2018 commented on GitHub (Feb 11, 2025): 同问，另外，基于满血版R1微调或者蒸馏具体需要什么配置？
bsakaguchi commented 2026-10-26 21:18:22 +00:00
Author
Owner
Copy link

@wingjson commented on GitHub (Feb 12, 2025):

同问，另外，基于满血版R1微调或者蒸馏具体需要什么配置？

满血部署都得1个t显存。微调至少x2

<!-- gh-comment-id:2653943108 --> @wingjson commented on GitHub (Feb 12, 2025): > 同问，另外，基于满血版R1微调或者蒸馏具体需要什么配置？ 满血部署都得1个t显存。微调至少x2
bsakaguchi commented 2026-10-26 21:18:22 +00:00
Author
Owner
Copy link

@alexhmyang commented on GitHub (Feb 13, 2025):

https://github.com/unslothai/unsloth/issues/1668#issuecomment-2653262963

unsloth 微调效果不对，完全不可用，上面是 测试过程

<!-- gh-comment-id:2656101807 --> @alexhmyang commented on GitHub (Feb 13, 2025): https://github.com/unslothai/unsloth/issues/1668#issuecomment-2653262963 unsloth 微调效果不对，完全不可用，上面是 测试过程
bsakaguchi commented 2026-10-26 21:18:22 +00:00
Author
Owner
Copy link

@github-actions[bot] commented on GitHub (Mar 25, 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:2749699114 --> @github-actions[bot] commented on GitHub (Mar 25, 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 21:18:23 +00:00
Author
Owner
Copy link

@github-actions[bot] commented on GitHub (Apr 30, 2025):

false

<!-- gh-comment-id:2840499938 --> @github-actions[bot] commented on GitHub (Apr 30, 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#269
No description provided.