Home > News content

Lesser Fullness of Grain opened up the first 100 billion parameter financial model "Xuanyuan" in China.

via:新浪科技     time:2023/5/27 9:00:23     readed:74

Sina Science and Technology News May 26th noon news, recently, Lesser Fullness of Grain officially opened up the country's first 100 billion-level Chinese financial model-"Xuanyuan". The Xuanyuan large model is trained on the basis of the Bloom model with 176 billion parameters. In the tasks of financial term understanding, financial market comment, financial data analysis and financial news understanding, the effect of the Xuanyuan big model is greatly improved compared with the general large model.

BLOOM (BigScience Language Open-science Open-access Multilingual) was created in 2021 by more than 1000 volunteer researchers in a project called Big Science BigScience and was officially launched on July 12, 2022. BLOOM has 176 billion parameters (variables that determine how the input data is converted into output), slightly more than GPT-3 with 175 billion parameters. BLOOM has 1.61TB text in 46 natural languages and 13 programming languages. Compared with the 13 billion-parameter LLaMA (Large Language Model Meta AI) model published by Meta, the number of Bloom parameters is more dominant.

At present, the hundreds of billions of Xuanyuan model can be downloaded in Huggingface and is open to all financial institutions.

Xu Dongliang, CTO of Lesser Fullness of Grain, said that the Xuanyuan big model is trained by the financial data accumulated in longitude Lesser Fullness of Grain business scenarios, and its understanding of finance-related issues has more advantages than the general model. We open the capacity of large models to financial institutions, which will help to promote the application of large models in the financial industry, reduce the application threshold of large models, and improve the level of intelligence of the financial industry.

translate engine: Tencent

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments