Open-R1: a fully open reproduction of DeepSeek-R1 from huggingface

Posted by siegevjorn@reddit | LocalLLaMA | View on Reddit | 8 comments

Reply to Post

8 Comments

[-]

Pro-editor-1105@reddit

Is this a paper or is it an actual model?

[-]

They are trying to use the paper and recreate to model from scratch, and they are building it openly on github with all data and code. Repo: [https://github.com/huggingface/open-r1](https://github.com/huggingface/open-r1)

[-]

siegevjorn@reddit (OP)

My understanding is It's none. Apparently they are working on it.

[-]

Specter_Origin@reddit

This has been posted bunch of times

[-]

siegevjorn@reddit (OP)

My apologies. I thought they made it work. But no, it's just an idea that they are working on.

[-]

siegevjorn@reddit (OP)

I realized that hf haven't made it work. It's just their plan of direction that they're taking. So there js nothing new, and the same as it was posted several days ago.

[-]

vasileer@reddit

was already posted

[-]

siegevjorn@reddit (OP)

Ok thanks