Taiwan-based Foxconn, the world’s largest contract electronics manufacturer, has unveiled its first large language model (LLM), named FoxBrain. The AI model aims to enhance manufacturing and supply chain management through data-driven intelligence and automation.
About FoxBrain & Its Capabilities
- FoxBrain is Taiwan’s first LLM with advanced reasoning capabilities, specifically optimized for traditional Chinese and Taiwanese language patterns.
- FoxBrain was trained on 120 of Nvidia’s H100 GPUs and completed training in just four weeks.
- Developed using Meta’s Llama 3.1 architecture and optimized for traditional Chinese and Taiwanese language styles.
- The model is slightly behind China’s DeepSeek distillation model in performance but is close to world-class standards.
Core Functions of FoxBrain
- Data Analysis & Decision Support
- Document Collaboration
- Mathematics, Reasoning & Problem-Solving
- Code Generation for Automation
Initial Applications:
- Primarily developed for internal use within Foxconn’s manufacturing and supply chain operations.
- Aims to enhance efficiency, intelligent decision-making, and production automation.
Nvidia’s Role & Supercomputing Support
- Nvidia’s Taipei-1 supercomputer, the largest in Taiwan, provided computing support and technical consultation during training. It is operated in Kaohsiung, a southern city on the island.
- Nvidia provided AI expertise and computational resources from its Taiwan-based data center.
- Foxconn, which also produces Nvidia AI servers, leveraged its partnership to ensure high-performance AI development.
Future Plans & Expansion
- Collaboration with technology partners to further expand FoxBrain’s applications in AI-driven manufacturing and supply chain management.
- Open-source sharing of selected AI models to promote AI development and adoption in the industry.
- Enhancing AI capabilities for global electronics manufacturing to improve efficiency, automation, and decision-making.