AFS Model Optimization Inference Service
Provides a one-stop solution for No-Code inference and optimization. Choose between Traditional Chinese enhancement and open-source models to support AI applications in enterprises and the public sector.
AFS Model Optimization Inference Service
Taizhi Cloud has launched an enterprise-level generative AI solution based on the Formosa Large Model (FFM), including "AFS ModelSpace (Model as a Service)" and "AFS Platform + AFS Cloud (One-stop large language model optimization and deployment service)". Inference and optimization can be completed in one stop without code. The AFS model optimization inference service supports various traditional Chinese enhanced models such as Llama3.3-FFM-70B-32K and FFM-Mixtral-8x7B, and also provides various open source models such as Meta Llama3.3-70B and the TAIDE model with Taiwan local characteristics. Whether you want to directly use various FFM and open-source large models, or build enterprise-specific models, AFS can provide the best solutions. Low start-up threshold and cost, saving the huge costs of building your own AIHPC, accelerating the development and innovative implementation of AI 2.0 applications!
AFS One-Stop Enterprise-Level Modeling Platform
AFS ModelSpace
Model-as-a-Service
Multiple Traditional Chinese enhanced FFM and mainstream open-source models such as Llama 3.3 and TAIDE are available, with dual-mode billing directly selectable.
✓ Easy to operateIt enables rapid deployment and launch of native large language model AI services in the cloud.
✓ Cost transparencyPay only for what you use; the pricing method is simple and easy to assess and control costs.
✓ Cloud HostingAutomation tools and management services simplify the management of large language models.
✓ Dual-mode billingBilling is based on the number of tokens used or the time spent building the model, choose as needed.
✓ User InterfaceBoth Public and Private Modes support Chat/Playground interfaces, making it easy to verify the latest models..
The new Public Mode chat interface is now available.
| AFS ModelSpace | Public Mode Billing based on the number of tokens used | Private Mode Billing based on model building time |
|---|---|---|
| Enhanced Feature Model (FFM) for Traditional Chinese | ⭐Llama 3.3-FFM ⭐Llama3.2-FFM Llama 3.1-FFM ⭐FFM-embedding-v2.1 FFM-embedding-v2 | Llama 3.3-FFM Llama 3.1-FFM Llama3-FFM FFM-Mistral FFM-Llama2-v2 FFM-Llama2 FFM-embedding-v2.1 FFM-embedding-v2 FFM-embedding |
| native model | TAIDE Meta-Llama 3.3 | TAIDE Meta-Llama 3.3 Meta-Llama 3.1 Meta-Llama2 Meta-CodeLlama |
*For complete product information and pricing, please refer to our website.AFS ModelSpaceProduct Price List
Trustworthy AI
Dialogue engine
The text data used to train the TAIDE large model is legally obtained and authorized, covering data from various government ministries and private sectors. The content is Taiwanese in nature and characteristics, responding to localization needs, ensuring the trustworthiness and applicability of generative AI, and helping to improve the AI development environment in Taiwan.
Cybersecurity Protection
Safety and compliance
The supercomputer is located in Taiwan and uses a Class A IDC data center service with national information and communication security responsibility level. It meets the highest information security assessment standards such as ISO, HIPPA, GDPR information security certification and auditability, as well as a high-quality SLA service quality commitment to ensure the secure processing and storage of sensitive data and support 24-hour customer service.
Quick Selection
Easy to get started
No programming language skills or additional downloads and installations are required. Simply create an account with one click to start using it. It offers an intuitive No-Code platform that is easy to use and quick to learn. It also features complete team management functions and an AI Chat interface for real-time interaction, effectively assisting in work and improving efficiency!
Applicable to multiple
Application areas
TAIDE features five main functions: automatic summarization, letter writing, article writing, Chinese-to-English and English-to-Chinese translation. It can assist in the processing and editing of official documents, reports, and proposals, saving tedious data, information, and research collection and providing supplementary analysis of policies, regulations, and other knowledge. Users can enjoy efficient AI services in their most familiar language, simplifying daily tasks and improving administrative efficiency and work quality.
AFS Platform
AFS Cloud
One-stop large language model optimization and deployment service
Provides enterprises with a one-stop cloud-based solution for quickly building their own proprietary large language models, from training to deployment.
✓ Reduce training costsIt provides FFM pre-trained models, so you don't have to start from scratch. You only need to prepare the data to start training your own model.
✓ Lower the barrier to entryNo need to build your own AI supercomputer; you can start training large language models with just a few thousand to tens of thousands of yuan.
✓ Simple and fastOne-stop service, enabling rapid deployment of optimized large language model AI services in the cloud, convenient and easy to use!
✓ Two optimization strategiesWe offer full parameter optimization and LoRA-efficient optimization strategies, allowing you to choose the most suitable strategy to create your custom model!
Llama 3.1-FFM
Llama3-FFM
FFM-Mistral
FFM-Llama2-v2
FFM-Llama2
Meta-Llama2
AFS Platform
Large Language Model Optimization Service
Highly flexible, meeting the customized model optimization needs of enterprises.
★ Suitable for full parameter optimization: Includes a large amount of complete training data and deep optimization capabilities.
★ LoRA optimization is suitable for: high-efficiency optimization with limited training data.
AFS Cloud
Large Language Model Cloud Deployment Service
After optimization, quickly deploy enterprise-specific large-scale models in the cloud.
★ Dedicated models for personal use: No need to share models, ensuring the security of sensitive information.
★ Chat for instant interaction: Start interacting with the model immediately using the Chat interface.
*For complete product information and pricing, please refer to our website.AFS PlatformandAFS CloudProduct Price List
Enterprise-level large model optimization solution
AFS Shuttle
Large-scale language model optimization for ride-sharing services
Low cost and rapid deployment: the best choice for proof of concept (POC).
Product Features
✓ Easy-to-use No-code Portal
✓ A variety of pre-trained models are available.
✓ Schedule programs based on customer needs and the total amount of readable tokens.
✓ National-level cybersecurity protection and tenant environment isolation
✓ It can be further integrated with secure network environments (including VPN).
✓ Offer a package at a fixed price to enterprise users to quickly obtain GPU resources for POC verification.
"AFS (AI Foundry Service)" provides a one-stop integrated service to help enterprises develop their own enterprise-grade generative AI models.
Fill out the form now for consultation
Free Consultation Service
Contact Taiwan AI Cloud experts to learn about and start using the solution that suits you.