Nubia launches self-developed Nebula large model: programming performance ranks first in China

According to news on July 23, this afternoon, Nubia held the Nubia AI+ dual flagship new product launch conference and launched the self-developed Nebula large model. In recent years, AI technology has experienced a blowout. Nubia has defined the product strategy and concept of "Al+", worked closely with the AI ​​large model R&D team, and launched the self-developed Nebula large model, which provides a natural language model. A series of highly professional AI large models such as models and visual large models, the model capabilities are far ahead in the industry.

1. In terms of programming performance, the Nebula large model ranks first in China and second only to GPT in the world.

2. With hundreds of millions of parameters, it can adapt to various usage scenarios from simple to complex.
3. Incremental pre-training is performed based on the self-developed large-scale Star Ring model. Domain data, Know-How knowledge accumulation, technical documents in the communication field, and wireless/core network/cloud code corpus are injected into the large model, using a parallel training framework.
4. Therefore, Nebula Large Model has strong knowledge and professional skills in the communication field and can adapt to different software development scenarios and needs.

   

The above is the detailed content of Nubia launches self-developed Nebula large model: programming performance ranks first in China. For more information, please follow other related articles on the PHP Chinese website!