News for hackers. Stuff that matters.

A comment on DeepSeek-R1 with Dynamic 1.58-bit Quantization. Parent comment

Re:DeepSeek-R1 with Dynamic 1.58-bit Quantization by Mashimo on Tuesday January 28, 2025 @07:59AM (#42851788)
"Qwen2.5 is the large language model series developed by Qwen team, Alibaba Cloud."

And I think they, the DeepSeek team, finetunes Qwen 7b on DeepSeek. That is how I understood it.

Which apparently makes it quite good for a 7b model. But, again: if I understood it correctly, is still just qween and without the reasoning of DeepSeek.

[ Parent | on HN ]
Re:DeepSeek-R1 with Dynamic 1.58-bit Quantization by randomifcpfan on Tuesday January 28, 2025 @09:10AM (#42852456)
In my application, code generation, the distilled DeepSeek models (7B to 70B) perform poorly. They imitate the reasoning of the r1 model, but their conclusions are not correct.

The real r1 model is great, better than o1, but the distilled models are not even as good as the base models that they were distilled from.

[ Parent | on HN ]