Wu Hanzhang, General Manager of ASUS Cloud and Taizhi Cloud, stated that Taizhi Cloud provides large language models and tool platforms in an open model, significantly reducing the entry barrier for enterprises and realizing inclusive AI and developing human-centered smart technology.
As the AI 2.0 wave sweeps the world, Taizhi Cloud launched the "Formosa Large Model" (FFM-176B and FFM-7B) based on the open-source large language model BLOOM at the Taiwan AI Supercomputing Conference in May. This time, it also launched the "Traditional Chinese Enhanced Version FFM-Llama 2 Large Language Models" globally for the first time, including the full series of 70B, 13B and 7B. It supports Code Llama program development and Embedding-enhanced semantic search models, and provides diverse large language models in cloud, ground and AI PC and complete AFS (AI Foundry Service) services to accelerate the application of AI 2.0 in enterprises.
Wu Han-chang, General Manager of ASUS Cloud and Taiwan AI Cloud, stated, "Over the past 30 years, Taiwan's ICT industry, based on an open architecture, has leaped to become a key global partner in digital innovation, with a flourishing of industries in chips, semiconductors, systems, and information services. Taiwan AI Cloud provides a reliable large language model through an open model, accumulating AI supercomputing capabilities and a complete AI tool platform, lowering the entry barrier for enterprises and accelerating the application of AI 2.0; we look forward to working with our partners to continue Taiwan's industrial advantages, realize inclusive AI, and develop human-centered smart technologies."
Based on the latest large language model Meta Llama 2, Taizhi Cloud's FFM-Llama 2 leverages AIHPC supercomputer computing power, a high-efficiency parallel computing environment, and large language model segmentation technology to train a large amount of local Taiwanese language data, significantly improving FFM-Llama 2's Traditional Chinese capabilities. The entire series of models can be fine-tuned, deployed, and used for inference. Enterprise users can also customize system prompts to make the AI model think more locally and respond to enterprise needs more accurately.
Taiwan AFS has launched a full-parameter and high-efficiency model fine-tuning (PEFT) function, which provides an affordable cloud or ground model inference solution optimized by GPTQ 4-bit high-performance compression technology. Since its release, the "Formosa Big Model" has served more than 70 large enterprises. Taiwan Mobile, a telecommunications operator, has adopted intelligent customer service into generative AI applications, and it is expected to expand to different industries within the group in the future to improve customer service efficiency.
By leveraging four major generative AI computing platforms from Taizhi Cloud—AIHPC Supercomputer Service, CCS (Container Compute Service) Container Computing Service with Built-in Generative AI Image Files, AFS (AI Foundry Service) Enterprise-Grade Generative AI Solution, and OneAI Machine Learning MLOps nocode AI Platform—Taizhi Cloud comprehensively meets the industry's needs for reliable, controllable, data-secure, and on-premises deployment of generative AI.
Taizhi Cloud has launched the world's first Traditional Chinese FFM-Llama 2 language model, which retains the excellent dialogue quality of the original Llama 2 and uses localized Chinese for accurate expression.
For more information, please see:https://tws.twcc.ai/by-theme/generative-ai-solution