Platform Introduction
PrimeHub is a Kubernetes-based platform designed for groups of data scientists. It aims at being a all-in-one enterprise machine learning platform to provide seamless MLOps experience.
Adopting the group-centric design, scientists can share datasets, artifacts and collaborates easily within same groups/projects, which accelerates collaborative development for project groups, besides, scientists are capable of developing models, serving models and monitoring models performance with full governance and transparency. In terms of platform administration, PrimeHub provides administrators capability of access control management, resources management and quotas control for project groups accordingly, which facilitates the efficiency of resources utilization.
Key Capabilities
- Cluster Computing
- One-Click Research Environments
- Easy Dataset Loading
- Management of Resource & Quota Privileges
- Custom Deep Learning Environments
- Enterprise-Class Account Management
Tiers
PrimeHub has PrimeHub Community, PrimeHub Enterprise and PrimeHub Deploy, three tiers. Regarding the differences, see Tiers Feature Comparison.
PrimeHub platform is composed of User Portal and Admin Portal; former provides dedicated features to data scientists, latter provides dedicated features to platform administrators.
User Portal
incorporates data-scientists-facing prominent features that scientists are able to turn workflows into automating pipelines by Job submission/schedule, able to prepare data/develop trained models from Notebook and able to deploy container-wrapped models as services by Model Deployment.
Getting Started as Users
Install the first 3rd-party application
NEW
ALPHA
As Group Admin
Admin Portal
incorporates administration features that administrators are able to do access-control management, resources/quotas control management, and to oversight the usage, moreover, administrators are able to build custom environments by customization of images.