Models & resources
讯飞星辰MaaS
An end-to-end model fine-tuning and deployment platform by iFlytek, connecting d
What is 讯飞星辰MaaS?
Xunfei Xingchen MaaS is an intelligent model fine-tuning service platform launched by iFlytek for developers.It builds an engineering solution that spans the entire 'data-model-service' link, aiming to reduce development costs and technical barriers for enterprise large model businesses.
The platform is compatible with mainstream open-source model ecosystems and supports third-party model hosting.Based on self-developed distributed training architecture and intelligent computing power scheduling systems, combined with efficient parameter optimization technologies like LoRA, it achieves standardization and engineering of large model customized development, accelerating industry intelligence upgrades.
Key features
- End-to-End EnginProvides end-to-end development capabilities including data enhancement, model fine-tuning, effect quantification, and one-click deployment.
- Efficient OptimiCombines efficient parameter optimization technologies like LoRA to improve customization efficiency and reduce computing consumption.
- Smart SchedulingBased on self-developed distributed training architecture and intelligent computing power scheduling systems to ensure efficient task execution.
- Open EcosystemCompatible with mainstream open-source model ecosystems and supports hosting and management of third-party models.
- Standardized DevAchieves standardization and engineering of large model customized development, simplifying complex technical processes.
How to use it
- 1.Visit the Xunfei Xingchen MaaS official website and register or log in.
- 2.Select a base model from the model square or upload/host a third-party model.
- 3.Use platform tools for data enhancement and preparation.
- 4.Configure fine-tuning parameters and optimize model training using technologies like LoRA.
- 5.Verify model performance through the effect quantification evaluation module.
- 6.Deploy the model as an API service with one click and integrate it into your business application.
Where it fits
- Enterprise CustoFine-tune general large models with private data to create exclusive AI assistants tailored to specific business logic.
- Open Source ModeProvide hosting environments for mainstream open-source models, quickly converting them into callable production-grade services.
- Industry IntelliHelp industries like finance and healthcare lower AI development barriers and accelerate the implementation of vertical large model applications.
Pricing and free quota
Pricing and any free quota follow the official site.
Who it is for
- AI Application Developers
- Enterprise Technical Teams
- Data Scientists
- Researchers needing model fine-tuning
FAQ
Which models does Xunfei Xingchen MaaS support?
The platform is compatible with mainstream open-source model ecosystems and supports third-party model hosting.Please check the official model square for the specific list of supported models.
What is LoRA technology?
LoRA is an efficient parameter optimization technology.Xunfei Xingchen MaaS combines this technology to help users achieve large model customization at a lower cost.
Does the platform provide data preprocessing features?
Yes, the platform provides data enhancement processing functions as part of the 'data-model-service' full-link solution.