NVIDIA has unveiled four new NVIDIA NIM microservices designed to empower developers in Japan and Taiwan to create high-performing generative AI applications tailored to local needs.
These microservices support the global push for sovereign AI, enabling nations to develop AI systems that align with regional values, laws, and interests.
Regional Focus: Japan and Taiwan
In Japan, NVIDIA’s Llama-3-Swallow-70B, trained on Japanese data, is enhancing AI applications by offering a deeper understanding of local regulations, customs, and language. The Tokyo Institute of Technology has fine-tuned this model to support various industries, including healthcare, where Preferred Networks is using it to develop a top-performing medical AI model.
In Taiwan, Llama-3-Taiwan-70B, based on Mandarin data, is being utilized by Chang Gung Memorial Hospital to improve medical communication and efficiency. This model is also being adopted by major companies like Pegatron to streamline operations and boost AI-driven outcomes across different sectors.
Enhancing AI with Local Expertise
NVIDIA’s NIM microservices are optimized for regional languages and cultural contexts, offering up to five times higher throughput and significantly lowering deployment costs. The RakutenAI 7B family of models, which are built on the Mistral-7B framework, has achieved leading performance in Japanese language tasks, further emphasizing the importance of region-specific AI training.
Supporting Sovereign AI Development
NVIDIA’s AI Foundry platform complements these efforts by enabling businesses in Japan and Taiwan to create custom AI models tailored to their unique needs. With tools for fine-tuning and a secure environment for deployment, NVIDIA is at the forefront of helping these nations develop AI that is both powerful and culturally relevant.
As Japan and Taiwan continue to invest in sovereign AI, NVIDIA’s new microservices provide the essential infrastructure to support their goals, ensuring that AI development remains closely aligned with local values and interests.
