Foxconn has unveiled a cutting-edge Large Language Model (LLM) called FoxBrain, initially designed for internal use. This FoxBrain model boasts capabilities in data analysis, mathematics, reasoning, problem-solving, and coding, operating in standard Mandarin Chinese.
Developed by the Hon Hai Research Institute, a research and development unit under Foxconn, FoxBrain was trained with 120 NVIDIA H100 GPUs in just 4 weeks, a notably brief timeframe with low training costs according to Foxconn. The model is based on Meta’s Llama 3.1 foundation, featuring 70 billion parameters, outperforming similar models like Llama-3-Taiwan-70B, especially in mathematics. However, Foxconn acknowledges that FoxBrain still lags behind the leading DeepSeek model, although it performs well on a global scale.
Foxconn plans to disseminate the FoxBrain model through various partners and open-source channels to advance AI usage in manufacturing and supply chain industries.
Source: Hon Hai via The Wall Street Journal
TLDR: Foxconn introduces FoxBrain, a sophisticated Large Language Model with advanced capabilities, developed internally and poised to enhance AI applications in manufacturing and supply chain sectors.
Leave a Comment