- HyperGAI, a pioneer in generative AI, develops multimodal HPT models capable of processing diverse data inputs.
- Collaborated with SMC’s H100 GPU clusters in Singapore to enhance computational performance and scalability.
- Utilized SMC’s high-performance network file system based on WEKA technology for efficient multimodal model training.
- Achieved significant environmental impact reduction, saving 29 tonnes of CO2 per month compared to traditional cloud solutions.
- Released three advanced HPT models in 2024, including HPT Pro and HPT 1.5 Air, showcasing superior performance in industry benchmarks.
Main AI News:
In a groundbreaking move towards sustainable AI innovation, HyperGAI, an applied AI company based in Singapore, has leveraged cutting-edge technology to train its Hyper Pre-Trained (HPT) models in collaboration with SMC, marking a significant milestone in multimodal AI development.
Founded with a mission to revolutionize generative AI technology, HyperGAI specializes in developing large language models (LLMs) capable of processing diverse inputs including text, images, videos, and more. Unlike conventional text-only LLMs, HyperGAI’s HPT framework represents a pioneering approach that facilitates the creation of versatile multimodal foundation models, enabling nuanced understanding and personalized content generation across various formats.
One of the initial challenges faced by HyperGAI was the development of foundation models adept at processing and generating personalized content from diverse data inputs. This necessitated a robust training infrastructure capable of handling intensive computational demands efficiently and sustainably.
In 2023, HyperGAI deployed SMC’s H100 GPU clusters within the Singapore cloud region, a strategic move that provided the company with the requisite GPU performance and scalability needed for its ambitious AI model development efforts. This collaboration enabled HyperGAI to achieve a significant milestone by releasing three HPT models within the first half of 2024 alone.
Key to this success was SMC’s state-of-the-art infrastructure, which includes a high-performance network file system based on WEKA technology. Designed specifically for handling large media files essential to multimodal model development, this system utilizes NVMe Flash to deliver unparalleled speed and scalability, crucial for meeting HyperGAI’s rigorous training requirements.
Moreover, by opting for SMC’s sustainable cloud solutions over traditional options, HyperGAI not only enhanced its computational capabilities but also significantly reduced its environmental footprint. The transition to SMC’s infrastructure reportedly led to a reduction of 29 tonnes of CO2 emissions per month, underscoring HyperGAI’s commitment to ethical and environmentally conscious AI innovation.
In March of this year, HyperGAI emerged from stealth mode with the launch of two flagship models: HPT Pro and HPT Air. These models showcase superior performance benchmarks compared to industry benchmarks like MMBench and SEED-Image, positioning them as formidable contenders in the AI landscape. In May, HyperGAI further strengthened its portfolio with the release of HPT 1.5 Air, featuring enhanced visual encoding capabilities and leveraging the latest advancements in the LLaMA 3 8B model.
With these innovations, HyperGAI has cemented its position at the forefront of multimodal AI research, setting new standards for performance and sustainability in the industry. By harnessing SMC’s advanced computational infrastructure, HyperGAI has not only accelerated its research and development efforts but has also established a blueprint for ethical and environmentally responsible AI innovation.
Conclusion:
HyperGAI’s strategic partnership with SMC not only propels the company to the forefront of multimodal AI development but also sets a precedent for sustainable innovation in the AI market. By leveraging SMC’s advanced infrastructure, HyperGAI has not only achieved remarkable performance milestones but has also demonstrated a commitment to reducing environmental impact—a critical consideration in today’s tech landscape. This collaboration underscores the growing importance of integrating sustainability into AI research and development, positioning HyperGAI as a leader in ethical and efficient AI innovation.
