Kunlun Wanwei announced open source "Tiangong" Skywork-13B series large model, 0 threshold for commercial use

CTOnews.com October 30 news, Kunlun Wanwei today announced the open source 10 billion-level large language model "Tiangong" Skywork-13B series, and supporting the open source 600GB, 150B Tokens open source Chinese data set.

Kunlun Wanwei "Tiangong" Skywork-13B series currently includes two models with 13 billion parameters: Skywork-13B-Base model and Skywork-13B-Math model. The open source address of CTOnews.com is as follows:

Skywork-13B download address (Model Scope): https://modelscope.cn/organization/skywork

Skywork-13B download address (Github): https://github.com/SkyworkAI/Skywork

In addition to the open source model, Skywork-13B series of large models will also open source 600GB, 150B Tokens Chinese corpus data set Skypile/Chinese-Web-Text-150B, which claims to be one of the largest open source Chinese data sets.

At the same time, Kunlun Wanwei "Tiangong" Skywork-13B series model is about to be fully available for commercial use-developers do not need to apply for commercial use.

According to reports, the open source Skywork-13B series models surpassed the open source models such as LLaMA2-13B (data as of October 25) in several evaluation benchmarks such as CEVAL, CMMLU, MMLU, GSM8K, etc.

In the evaluation of the field of Chinese text creation, the achievements of Skywork-13B series of large models are as follows, and they perform well in the fields of science and technology, finance, government affairs, enterprise services, cultural creation, games and so on.