Editions

TigerGraph ML Workbench is available as a service through TigerGraph Cloud or standalone in two editions, Developer and Enterprise.

While both on-prem editions offer the same useful Python-level features for your data science needs, the Enterprise edition offers more powerful features to support production-level model training on enterprise-level data sets.

The tables below can help you compare the key differences between the editions.

ML Workbench on TigerGraph Cloud

ML Workbench on TigerGraph Cloud is built on KubeFlow, allowing it to offer several unique AutoML and visualization features conveniently in a single web framework.

TigerGraph’s Machine Learning Workbench is a fully-managed platform that accelerates development and deployment of graph-enhanced machine learning, which leverages the added insights from connected data and graph features for better predictions. ML Workbench delivers a scalable Graph ML model orchestrated by Kubeflow running on Kubernetes containers.

Due to its accurate predictive power stemming from unique graph features and graph models, the workbench enables organizations to unlock even better insights and greater business impact at scale.

Please see the official product page for early access information.

Developer and Enterprise On-Prem Editions

Features

Developer Enterprise

Compatibility

  • TigerGraph Database

  • Amazon SageMaker, Azure ML, GCP Vertex

  • PyG, DGL ML Framework

  • TigerGraph Database

  • Amazon SageMaker, Azure ML, GCP Vertex

  • PyG, DGL ML Framework

Onboarding

  • ML Workbench Docker Images

  • MacOS and Linux Installers

  • pip install & conda install

  • ML Workbench Docker Images

  • MacOS and Linux Installers

  • pip install & conda install

Capabilities

Python-level capabilities with pyTigerGraph:

  • Graph data partitioning

  • Graph Data Loading & Export (HTTP)

  • Subgraph sampling

  • Data Batching

  • Graph feature generation

  • GNN: Homogeneous Graph Support

  • GNN: Node Prediction support

  • GNN: Heterogeneous Graph Support

  • GNN: Link Prediction Support

  • GNN Inference with real-time data

Python-level capabilities with pyTigerGraph:

  • Graph data partitioning

  • Graph Data Loading & Export (HTTP & Kafka)

  • Subgraph sampling

  • Data Batching

  • Graph feature generation

  • GNN: Homogeneous Graph Support

  • GNN: Node Prediction support

  • GNN: Heterogeneous Graph Support

  • GNN: Link Prediction Support

  • GNN Inference with real-time data

Data Export Method

HTTP only

Reliable and efficient data export via both HTTP and Kafka

Data Export Size

Limited to 2GB

Unlimited

Parallel Training

No

Yes

Support

Community support

  • 10 hr Professional Service / Consulting on Solution Building

  • Standard support SLA with 12 x 5

Customer Scenarios

Developer Edition Enterprise Edition

Purpose

  • Learning Graph Databases

  • Learning Graph Data Science

  • Building proofs of concept

Production deployment of Graph ML Models

Audience

  • Students

  • Researchers

  • ML Practitioners

Enterprise data science teams

Infrastructure Readiness

Local machines / Internal ML infra (self-managed)

Local machines / Internal ML Infra (self-managed)

Data Size

Small data set (<2GB)

Production / Enterprise level data

Performance requirements

  • Low requirements for training time

  • Low requirements for model accuracy / predictability

  • High requirements for training time (near real time)

  • High requirements for model accuracy / predictability