[GH-ISSUE #419] r1和v3 哪个先出现呢? #330

Closed
opened 2026-10-26 21:19:31 +00:00 by bsakaguchi · 3 comments
Owner

Originally created by @carter54 on GitHub (Feb 16, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-R1/issues/419

请教一个问题哈

r1论文里写sft训练数据中有20w通用数据从v3生成,但是v3论文里又写v3训练数据里推理数据是来自r1

那是先有r1还是先有v3呢? 两个模型是否是互相迭代训练优化的?

Originally created by @carter54 on GitHub (Feb 16, 2025). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-R1/issues/419 请教一个问题哈 r1论文里写sft训练数据中有20w通用数据从v3生成,但是v3论文里又写v3训练数据里推理数据是来自r1 那是先有r1还是先有v3呢? 两个模型是否是互相迭代训练优化的?
Author
Owner

@yijianguanzhu commented on GitHub (Feb 17, 2025):

先V3后R1,R1推出后才爆火的

<!-- gh-comment-id:2662369850 --> @yijianguanzhu commented on GitHub (Feb 17, 2025): 先V3后R1,R1推出后才爆火的
Author
Owner

@ictzyqq commented on GitHub (Feb 17, 2025):

我也有相同的疑问

<!-- gh-comment-id:2662554030 --> @ictzyqq commented on GitHub (Feb 17, 2025): 我也有相同的疑问
Author
Owner

@hhnqqq commented on GitHub (Feb 19, 2025):

左脚踩右脚上天

<!-- gh-comment-id:2667478212 --> @hhnqqq commented on GitHub (Feb 19, 2025): 左脚踩右脚上天
Sign in to join this conversation.
No milestone
No project
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#330
No description provided.