Foxconn launches first major language model

Taiwanese company Foxconn announced today (March 10) that it has developed the first large-scale language model and plans to use the technology to improve manufacturing and supply chain management.

The model — dubbed “FoxBrain” — was trained using 120 of Nvidia’s H100 GPUs and completed in about four weeks, the world’s largest contract chipmaker said.

Foxconn launches first major language model, very close to world-leading standards.

The company, which assembles iPhones for Apple and also makes Nvidia's artificial intelligence servers, said the model is based on Meta's Llama 3.1 architecture.

This is Taiwan's first large language model with inference capabilities, optimized for Traditional Chinese and Taiwanese language styles.

Foxconn said that although there is a small performance gap compared to DeepSeek's distilled model from China, FoxBrain's overall performance is very close to world-leading standards.

Originally designed for internal applications, FoxBrain includes data analysis, decision support, document collaboration, mathematics, reasoning and problem solving, and programming code generation.

Foxconn plans to work with technology partners to expand the application of the model, share open source information and leverage artificial intelligence in manufacturing, supply chain management and smart decision making.

According to Foxconn, Nvidia provided support through the "Taipei-1" supercomputer located in Taiwan and provided technical advice during model training.

Taipei-1, the largest supercomputer in Taiwan, is owned and operated by Nvidia in Kaohsiung, a city in the south of the island.

Foxconn is expected to announce more details about the model at Nvidia's GTC developer conference in mid-March.

(Source Reuters)

Source: https://www.baogiaothong.vn/foxconn-ra-mat-mo-hinh-ngon-ngu-lon-dau-tien-192250310181358802.htm