February 24, AI Large Modelcommunal Hugging Face Co-founder and CEO Clement Delangue announced in a post early this morning:DeepSeek-R1 won at Hugging Face.Over 10,000 LikesThe platform's modeling capabilities have been recognized as one of the platform's nearly 1.5 million models.Most Popular Large Models, "It's shocking!"
Late last month, Hugging Face also announced that Open-R1 projectThe company plans to fill in some of the components of DeepSeek R1 that are not open source and "add more fuel to the fire" of this global boom.
According to Hugging Face, while the weights used by DeepSeek are known, the dataset and code used to train the model are unknown.With Open-R1, Hugging Face hopes to fill these gaps.
As 1AI previously reported, Hugging Face called the work important because DeepSeek R1 is so efficient thatCan be used as a base model for innovation. It can also be used by researchers, scientists and businesses as an affordable model for promoting innovation and breakthroughs.
Hugging Face says the synthetic dataset it plans to build will not only focus on math, but will also explore other areas for other fields such as science.
