Hacker News Reader

Comment by dotancohen | Hacker News Reader

Not many non-gamers have hardware capable of running such a model locally - never mind the skills.

For most people, bash is not a tool for interacting with the computer, it is how they express their frustration with the computer (sometimes leaving damaged keyboards).

View on HN

I have DeepSeek-R1 1.5b running on a Raspberry Pi 5. I have DS-R1 14b Q6 running on my old AM4 Ryzen with a AMD GPU, without issues. My primary workstation is running 32B Q8 and without issues. And it's simple!

That's not the DeepSeek R1 model that they're offering via the API on these servers. That's a Qwen model that's been fine-tuned on output from the big R1 model.

Source?

https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-...

DeepSeek-R1-Distill-Qwen-1.5B, DeepSeek-R1-Distill-Qwen-7B, DeepSeek-R1-Distill-Qwen-14B and DeepSeek-R1-Distill-Qwen-32B are derived from Qwen-2.5 series, which are originally licensed under Apache 2.0 License, and now finetuned with 800k samples curated with DeepSeek-R1. DeepSeek-R1-Distill-Llama-8B is derived from Llama3.1-8B-Base and is originally licensed under llama3.1 license. DeepSeek-R1-Distill-Llama-70B is derived from Llama3.3-70B-Instruct and is originally licensed under llama3.3 license.

https://arxiv.org/abs/2501.12948

Wow all the gamers with mad LLM skillz.

Pretty sure gamers are mentioned because those are the usual demo that has GPUs with enough memory outside of people in the ML industry.

So you’re in the demo scene as well? Yay

You were not able to use context clues to figure that "demo" in this case is short for "demographics"? Sad

Reply on news.ycombinator.com