open-r1

One of my favorite code comments of all time is still in the src: "# TODO: implement a proper validator to compare against ground truth. For now we just check for exact string match on each line of stdout." [1] This was one of my chief complaints about the entire R1 news cycle, it felt like no one actually read the technical report. They were being heralded for their openness, but they left out the most meaningful details that you'd need to reproduce their work. [1] https://github.com/huggingface/open-r1/blob/1416fa0cf21595d2...

It’s all open source and even their methods are published. Berkeley could replicate the reasoning principle of R1 with 30$ compute budget. Open-R1 aims to fully replicate R1 results with published methods and recipes. Their distill results look already very impressive. All these open source models are based on Meta Llama and open to everyone. Why should western labs and universities not be able to continue and innovate with open source models? I don’t see why we have to rely on China. Keeping the open source projects open is however extremely important. And for that we should fight. Not chasing conspiracy theories or political narratives. https://github.com/huggingface/open-r1

I used to think DeepSeek was also censored because of the system prompt but that was not the case, it was inherent in it's training. It's the same reason HuggingFace and Perplexity trained their own DeepSeek (Open-r1[0] and r1-1776[1]) instead of just changing the system prompt. There's no doubt that Grok will go the same way. They tried tweaking it with system prompts and got caught so this is the next step. 0. https://github.com/huggingface/open-r1 1. https://playground.perplexity.ai/

No sign of what source material it was trained on though right? So open weight rather than reproducible from source. I remember there's a project "Open R1" that last I checked was working on gathering their own list of training material, looks active but not sure how far along they've gotten: https://github.com/huggingface/open-r1

Kagi may be using the official DeepSeek API and not hosting the model itself. There is work being done to make it completely uncensored: https://github.com/huggingface/open-r1 https://ollama.com/huihui_ai/deepseek-r1-abliterated I was mistaken though, it is more than just a system prompt causing the censorship.

So does this mean we will not be able to download works like open-r1, an open reproduction of deepseek? https://github.com/huggingface/open-r1

> The model isn’t reproducible Not necessarily[0], it's a WIP, but: https://github.com/huggingface/open-r1 [0] Surely they won't end up with the exact same weights, but it should be possible to verify something about the model and approach

Huggingface is working on reproducing it: https://github.com/huggingface/open-r1

Deepseek R1 is the closest thing we have to fully open-source currently. Open enough that Huggingface is recreating R1 completely out in the open. https://github.com/huggingface/open-r1

Huggingface is working on reproducing R1, as I'm sure many others are https://github.com/huggingface/open-r1

They seem to have matched OpenAI model capability on a fraction of the resources. R1 is roughly as good as o1 and can be run locally. There are some interesting contributions in the paper too. So I see it as two parts, (1) the money, lots invested, wondering if they will ever make it back now (2) China showing impressive results while being handicapped by the West https://arxiv.org/abs/2501.12948 Keep an eye on the effort to reproduce here: https://github.com/huggingface/open-r1 We will see if the (over?) reaction matches reality in time. Media sure loves to whipsaw us all around