You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
thomas 18a76ed026 upload english user manual 10 months ago
.github/workflows rm un-used action yaml 1 year ago
devenv merge from master 1 year ago
devenv.arm64 merge from master 1 year ago
dlworkspace-compile fix comflict 10 months ago
docs upload english user manual 10 months ago
src fix comflict 10 months ago
.gitattributes update docs 1 year ago
.gitignore fix comflict 10 months ago
.gitmodules add k8x-pom dependance in .gitmodules 1 year ago Update installing prerequisities, update document. 4 years ago update install doc 1 year ago
Jenkinsfile modify jenkinsfile image name 1 year ago
LICENSE update license 11 months ago
NOTICE update readme 10 months ago update license 10 months ago update license 10 months ago
azure-pipelines.yml init 1 year ago bugfix for docker env 1 year ago fix arm64 deploy 1 year ago merged from microsoft v1.5 1 year ago
version-info improve version example 1 year ago





Apulis AI Platform is designed to provide an end-to-end AI solution to users from different industries and enable them to start the high-performance AI development work with significantly reduced ramp up time, thereby saving costs and improving efficiency. It will also provide a highly efficient, low cost private cloud AI solution for small and medium size company.

The platform incorporates TensorFlow, PyTorch, MindSpore and other open source AI frameworks, thereby provides user friendly development environment for AI model training, auto ML, hardware status monitoring etc., making it very easy for AI developers to quickly develop AI application. It also has built-in comprehensive early warning system which can automatically alert the system administrator on any anomaly, thereby improve the platform efficiency and security.

The platform adopts the lightweight virtualization technologies, such as Docker containers that containerizes one or more programs, and provide a standard management interface. Each container is separated from each other. Kubernetes clustering technology is used to orchestrate the containerized applications for planning, automated deployment, updates, and maintenance..

Directory Structure

|-- devenv                          Dockerfile for creating dev environment on amd64 arch
|-- devenv.arm64                    Dockerfile for creating dev environment on arm64 arch
|-- docs
|   |-- deployment
|   `-- tutorial
|-- example
|   `-- resnet50_cifar10
|-- License
`-- src
    |-- ARM
    |-- ClusterBootstrap            deployment module
    |-- ClusterManager              main backend module
    |-- ClusterPortal
    |-- Jobs_Templete               AI Job template
    |-- RepairManager               Alert module
    |-- RestAPI                     API module
    |-- StorageManager
    |-- WebUI                       Frontend module
    |-- WebUI2
    |-- dashboard
    |-- dev-utils		    
    |-- docker-images               Miscellaneus components          
    |-- init-scripts                Initialization scripts for AI jobs
    |-- user-dashboard
    |-- user-synchronizer
    `-- utils                       utility module


The entire codebase is under MIT license


依瞳人工智能平台旨在为不同行业的用户提供基于深度学习的端到端解决方案,使用户可以用最快的速度、最少的时间开始高性能的深度学习工作,从而大幅节省研究成本、提高研发效率,同时可为中小企业解决私有云难建成、成本高等问题。 平台融合了Tensorflow、PyTorch、MindSpore等开源深度学习框架,提供了模型训练、超参调优、集群状态监控等开发环境,方便AI开发者快速搭建人工智能开发环境,开展AI开发应用。在监控模块基础上搭建预警模块,自动将平台异常通知管理员,提升平台的预警效率及安全性能。

JavaScript C Python C++ YAML other