DOCSEARCHbyAlgolia
Submit Docs

AirCloud Documentation

website·Indexed 32 days ago

https://docs.aieev.cloud

A decentralized cloud computing platform designed for AI, machine learning, and various computing-based applications. It provides GPU resources ranging from NVIDIA RTX 3070 to RTX 6000 Ada Generation and features Air Container for deploying and managing custom container images.

Tokens
2.8K
Snippets
5
Records
27
Agent score
49%

What's inside AirCloud

  1. Overview of Air Cloud decentralized computing

    Air Cloud is a decentralized cloud platform designed for AI and machine learning applications. It allows users to deploy services and containerized applications using a variety of GPU and CPU resources. The platform supports serverless Auto Scaling to enable easy expansion of AI services without complex configuration.

  2. Introduction to Air Container

    Air Container is a platform that allows users to deploy and manage custom-built container images on Air Cloud. It enables the execution of AI models, applications, and services in a scalable, cloud-based environment that can be flexibly expanded to meet diverse user requirements.

  3. Deploy infrastructure using the Web-Based GUI Console

    The platform provides a web-based GUI console that allows users to deploy their infrastructure through a visual interface.

  4. Integrate AirCloud using SDKs

    AirCloud provides SDKs for various programming languages to facilitate integration and automation. Refer to the specific language SDK documentation for implementation details. (Note: This feature is listed as upcoming/planned).

  5. Import models or containers via Click-away Migration

    Use the Click-away Migration feature to import models or containers directly from a repository into the environment.

  6. Deploy a container via AirCloud Console

    To deploy a container, navigate to the Air Container menu in the AirCloud console and click the 'New Container Deployment' button. You will need to provide the image URL and configure the required resources (CPU, GPU, memory, and storage). Note that Auto Scaling is planned for future support.

  7. Create an AirCloud account

    To use AirCloud and manage its resources, you must create an account. Currently, account registration is only available via GitHub. Once registered, you can log in to access AirCloud services.

  8. Identify available GPU models in AirCloud

    AirCloud provides a variety of NVIDIA RTX series GPUs, ranging from the RTX 3070 to the RTX 6000 Ada Generation. Users can select a specific model based on their requirements for memory size and performance.

  9. View and manage Air Container logs

    Air Container allows users to monitor the output of containers distributed across multiple nodes. Logs are stored and separated by date, and users can view up to 50,000 lines of log data. Currently, logs displayed for each node are aggregated from multiple containers; individual container-level log filtering is a planned future feature. To access logs, navigate to the Logs menu in the AirCloud console to view the list of log files per node and inspect their detailed contents.

  12. Prepare container images for AirCloud deployment

    Before deploying to AirCloud, applications must be containerized using a platform like Docker and uploaded to a supported container registry. Currently supported registries include Docker Hub, AWS ECR, and GitHub Container Registry.