logo
Alauda AI
English
Русский
English
Русский
logo
Alauda AI
Navigation

Overview

Introduction
Quick Start
Release Notes

Install

Pre-installation Configuration
Install Alauda AI Essentials
Install Alauda AI

Upgrade

Upgrade from AI 1.3

Uninstall

Uninstall

Infrastructure Management

Device Management

About Alauda Build of Hami
About Alauda Build of NVIDIA GPU Device Plugin

Multi-Tenant

Guides

Namespace Management

Workbench

Overview

Introduction
Install
Upgrade

How To

Create WorkspaceKind
Create Workbench

Model Deployment & Inference

Overview

Introduction
Features

Inference Service

Introduction

Guides

Inference Service

How To

Extend Inference Runtimes
Configure External Access for Inference Services
Configure Scaling for Inference Services

Troubleshooting

Experiencing Inference Service Timeouts with MLServer Runtime
Inference Service Fails to Enter Running State

Model Management

Introduction

Guides

Model Repository

Monitoring & Ops

Overview

Introduction
Features Overview

Logging & Tracing

Introduction

Guides

Logging

Resource Monitoring

Introduction

Guides

Resource Monitoring

API Reference

Introduction

Kubernetes APIs

Inference Service APIs

ClusterServingRuntime [serving.kserve.io/v1alpha1]
InferenceService [serving.kserve.io/v1beta1]

Workbench APIs

Workspace Kind [kubeflow.org/v1beta1]
Workspace [kubeflow.org/v1beta1]

Manage APIs

AmlNamespace [manage.aml.dev/v1alpha1]

Operator APIs

AmlCluster [amlclusters.aml.dev/v1alpha1]
Glossary
Previous PageOverview
Next PageInstall

#Introduction

#TOC

#Workbench

Workbench manages web IDE instances like Jupyter Notebooks, code-server, or RStudio. AI/ML developers and data scientists can use cluster resources (like GPUs) and connect to in-cluster services to create general ML jobs and pipelines.

Workbench uses "Kubeflow Notebook 2.0" as its backend to create cloud IDE containers. If you are familiar with "Kubeflow Notebook", it's easy to start using Workbench.

Some key features include:

  • Native support for JupyterLab, Visual Studio Code (code-server), and RStudio (in the future).
  • Users can create notebook containers directly in the cluster, rather than locally on their workstations.
  • Admins can provide standard notebook images for their organization with required packages pre-installed.
  • Access control is managed by Kubeflow's RBAC, enabling easier notebook sharing across the organization.

#WorkspaceKind

  • this is the resource that cluster-admins create
  • it specifies the template for a Workspace (e.g. "JupyterLab", "VSCode", "RStudio")
  • initially, we would only support a "podTemplate" kind, which is very similar to the existing Notebook CRD, but in the future, we could support other types of templates (e.g. "helmTemplate")