[GH-ISSUE #8698] Rename the codename of DeepSeek-R1 fine-tuned models #5637

Closed
opened 2026-04-12 16:55:23 -05:00 by GiteaMirror · 2 comments
Owner

Originally created by @blacklightpy on GitHub (Jan 30, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8698

Only the deepseek-r1:671b should be under the deepseek-r1 codename for now. Because the fine-tuned models (termed as distills by DeepSeek AI) are not the same models, and are misleading.

All the others are distills, and should either come under something like deepseek-r1-distill, like deepseek-r1-distill:llama3.1-7b, deepseek-r1-distill:qwen2.5-1.5b.

Or they should be under the respective models as finetunes, like llama3.1:deepseek-r1-distill-7b, and qwen2.5:deepseek-r1-distill-1.5b.

Originally created by @blacklightpy on GitHub (Jan 30, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8698 Only the `deepseek-r1:671b` should be under the `deepseek-r1` codename for now. Because the fine-tuned models (termed as distills by DeepSeek AI) are not the same models, and are misleading. All the others are distills, and should either come under something like `deepseek-r1-distill`, like `deepseek-r1-distill:llama3.1-7b`, `deepseek-r1-distill:qwen2.5-1.5b`. Or they should be under the respective models as finetunes, like `llama3.1:deepseek-r1-distill-7b`, and `qwen2.5:deepseek-r1-distill-1.5b`.
Author
Owner

@philippstoboy commented on GitHub (Jan 30, 2025):

Sounds like a great idea! I believe your second approach would be very beneficial!

<!-- gh-comment-id:2625046445 --> @philippstoboy commented on GitHub (Jan 30, 2025): Sounds like a great idea! I believe your second approach would be very beneficial!
Author
Owner

@rick-github commented on GitHub (Jan 30, 2025):

https://github.com/ollama/ollama/issues/8557

<!-- gh-comment-id:2625049136 --> @rick-github commented on GitHub (Jan 30, 2025): https://github.com/ollama/ollama/issues/8557
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#5637