What is Lanrui Xingzhou
Lanrui Xingzhou is a cloud-based AI training and inference integrated computing power platform, independently developed by Yifang Jianshu. It provides cost-effective GPU computing resources, out-of-the-box training and inference environments, and AI toolchains, aiming to solve the security and trust issues in the AI collaboration process and accelerate the release of AI value. The platform supports multi-machine multi-GPU distributed training, provides high-availability APIs and an open image ecosystem, and has built-in launchers for various AI applications, such as ComfyUI, SD WebUI, etc., and supports functions such as voice conversion and speech synthesis. Lanrui Xingzhou provides a rich set of pre-trained models, datasets, and custom inference services, supporting developers in quickly learning and deploying large models.
Main features and services of Lanrui Xingzhou
- Inference service: Provides one-click deployment of production-grade inference services, supports users in uploading their own models, and supports models of multiple modalities and inference engine frameworks. It has functions such as asynchronous invocation and inference result callbacks, and supports persistence of inference results and CDN distribution.
- Workspace: Users can quickly set up a complete development environment, supporting tools such as JupyterLab, VS Code, and TensorBoard.
- Xingzhou API: Provides inference services for mainstream large models, simplifying the model deployment and maintenance process. Users can start using it with just a few lines of code.
- Inference community: Users can share application prototypes built with the Xingzhou API, and can also explore application prototypes shared by other users.
- Image community: Users can share images they have built, or explore images shared by other users.
- Application version and Professional version：
- The application version provides the latest and hottest AI applications, and users can use them online or deploy the AI application environment with one click.
- The professional version provides functions such as pre-trained models, datasets, inference services, and community images, and is suitable for developers with a certain development foundation.
- GPU rental: Provides a variety of GPU rental options, supports pay-as-you-go or time-period billing, helping users quickly deploy and run AI applications.
- AI toolchain: Integrates a variety of AI toolchains, supports one-stop MLOps solutions, and supports multi-machine multi-GPU distributed training.
- High-availability APIs and open image ecosystem: Built-in launchers for various AI applications, such as ComfyUI, SD WebUI, etc., supporting functions such as voice conversion and speech synthesis
Lanrui Xingzhou's product official website
- Product official website
- Old address：lanrui-ai.com(Data on this platform will not be retained after November 1, 2024; please migrate to the new platform as soon as possible)
- New address：lanrui.co
How to use Lanrui Xingzhou
- Registration and login: Visit the official website of Lanrui Xingzhou to register or log in as a user.
- Choose computing resources: In the platform's computing power marketplace, choose pay-as-you-go or weekly/monthly computing resources according to your needs.
- Create a workspace: Choose pre-trained models, datasets, and development environment images as needed to create a workspace. The workspace supports tools such as JupyterLab and VS Code.
- Application launcher: The platform provides an application launcher, supporting users in quickly deploying application images, such as popular AI painting applications like SD WebUI and ComfyUI.
- Custom creation: For users with specific needs, they can choose to create a custom workspace to flexibly configure the environment and storage.
- Using Xingzhou API: The platform provides the Xingzhou API service, supports inference services for mainstream large models, and users can call it through API interfaces to quickly integrate AI capabilities.
- Start development and training: In the workspace, use the computing power provided by the platform for AI model development and training.
- Monitoring and logs: Use features such as Grafana monitoring, event logs, and SSH login to manage and monitor your AI training tasks.
- Stop and billing management: After use, remember to stop the workspace to avoid unnecessary billing. The platform also supports setting an automatic stop time to better manage resources and costs.
Lanrui Xingzhou product pricing
- Computing resources：
- Lanrui Xingzhou provides a variety of GPU rental options, supporting pay-as-you-go billing and time-package billing (such as weekly/monthly/yearly packages).
- Pay-as-you-go billing is based on actual usage time, with the minimum billing period being minutes.
- At startup, fees are pre-deducted based on the minimum billing period of 1 hour, and billing starts when the status changes from starting to running
- After stopping, refunds are made at the minute level based on actual usage duration
- Less than 1 minute is billed as 1 minute
- Stopping in the starting state incurs no charge, and the fee is returned to the balance in 1-2 minutes
- Network disk service: Each user receives 10GB of storage space for free, and the excess will be billed based on actual usage,
- The billing formula is: fee = storage amount (GB) × monthly unit price ÷ 30 days ÷ 24 hours.
- Monthly unit price: 0.15 yuan/GB/month
- Data disk and environment persistent disk: The fees for the data disk and environment persistent disk are billed hourly, starting from creation and ending after destruction.
- Xingzhou API: The Xingzhou API service is paid per call, and the platform will charge based on the actual inference runtime cost.
- New user discount: Lanrui Xingzhou provides free computing power and a newcomer gift pack for newly registered users to encourage users to experience the platform's services.
- Arrears policy：
- When the user's account balance is less than 0, the workspace will be automatically shut down.
- Network disk data supports viewing/deleting files, but uploading/downloading files is not allowed.
- The data disk and environment persistent disk will continue to be billed during arrears. If arrears continue beyond a certain period, the data will be cleared or destroyed.
Lanrui Xingzhou application scenarios
- AI painting and creation: The platform provides AI painting applications such as SD WebUI and ComfyUI. Users can deploy these applications with one click through the application launcher for AI painting creation.
- Model training and inference: Lanrui Xingzhou supports AI development environment images for a variety of machine learning algorithms and frameworks. Users can choose pay-as-you-go or yearly/monthly billed GPU resources for model training according to their needs.
- Large model API service: The platform provides a rich set of large model AI applications that users can experience online, and can call per use or exclusively deploy these large model APIs.
- Privacy-preserving computing: Lanrui Xingzhou integrates privacy-preserving computing technology and provides trusted AI computing nodes, focusing on solving security and trust issues in the AI supply-demand market and promoting the realization of data value.
- Dataset and model sharing: Users can share application prototypes built with the Xingzhou API in the inference community, or explore application prototypes shared by other users. Users can also share or discover images shared by other users in the image community.
- Cloud-hosted offline tasks: The platform supports cloud-hosted offline tasks, greatly reducing task costs and improving resource efficiency.