[GH-ISSUE #8913] DeepSeek R1 #83464

Closed
opened 2026-05-09 18:17:00 -05:00 by GiteaMirror · 2 comments
Owner

Originally created by @20246688 on GitHub (Feb 7, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8913

Hello. I noticed that the 1.5b, 7b, and 8b models in DeepSeek R1 are about the same size as the pre-distilled Qwen 2.5 series. I’m wondering what the difference between them is?

Originally created by @20246688 on GitHub (Feb 7, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8913 Hello. I noticed that the 1.5b, 7b, and 8b models in DeepSeek R1 are about the same size as the pre-distilled Qwen 2.5 series. I’m wondering what the difference between them is?
GiteaMirror added the model label 2026-05-09 18:17:00 -05:00
Author
Owner

@gzzzzzd commented on GitHub (Feb 7, 2025):

You can check out the original paper. Except for DeepSeek-671B, other models are distill model by fine-tuning the Qwen series models with datas generating by DeepSeek-R1.

<!-- gh-comment-id:2641852636 --> @gzzzzzd commented on GitHub (Feb 7, 2025): You can check out the original paper. Except for DeepSeek-671B, other models are distill model by fine-tuning the Qwen series models with datas generating by DeepSeek-R1.
Author
Owner

@20246688 commented on GitHub (Feb 7, 2025):

您可以查看原始论文。除 DeepSeek-671B 外,其他模型都是通过使用 DeepSeek-R1 生成的数据对 Qwen 系列模型进行微调来提炼模型。

thanks a lot

<!-- gh-comment-id:2641857594 --> @20246688 commented on GitHub (Feb 7, 2025): > 您可以查看原始论文。除 DeepSeek-671B 外,其他模型都是通过使用 DeepSeek-R1 生成的数据对 Qwen 系列模型进行微调来提炼模型。 thanks a lot
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#83464