# awslabs/dgl-ke

High performance, easy-to-use, and scalable package for learning large-scale knowledge graph embeddings.

Repository: https://github.com/awslabs/dgl-ke
Canonical: https://ross.abutalabs.com/products/dgl-ke
Homepage: https://dglke.dgl.ai/doc/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: machine-learning, knowledge-graph, knowledge-graphs-embeddings, graph-learning, dgl
Last push: 2026-07-06T18:22:51+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 91, release rhythm 8, longevity 100
- inputs: {"age_days": 2375, "days_push": 58, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1331, forks 201 (observed 2026-08-28T04:04:24.340917+00:00)

## What it is
DGL-KE is a high-performance Python package built on Deep Graph Library (DGL) for training, evaluating, and inferring knowledge graph embeddings at scale. It supports popular models like TransE, TransR, RESCAL, DistMult, ComplEx, and RotatE, and scales to graphs with millions of nodes and billions of edges on CPUs, GPUs, or clusters.

## Use cases
- train knowledge graph embeddings on large graphs
- compute embeddings for a knowledge graph with billions of edges
- evaluate link prediction quality of KG embeddings
- predict missing entities or relations in a knowledge graph
- run distributed knowledge graph embedding training on a cluster
- find similar entities using embedding similarity

## When to choose
- you need to train KG embeddings on very large graphs (millions of nodes, billions of edges) with GPU or distributed support
- you want a CLI-driven workflow for training, evaluating, and predicting with standard KGE models like TransE, DistMult, ComplEx, or RotatE
- you are already using the Deep Graph Library ecosystem

## When to avoid
- you only need basic TransE, DistMult, or RotatE training - the AWS GraphStorm project is the recommended successor for those models
- you need general-purpose graph neural networks rather than knowledge graph embeddings
- you need a Python API-first library rather than CLI commands

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, cli, benchmarking
- domain: machine-learning, graph-processing, large-language-models, data-science
- platform: python, cloud
- tags: knowledge-graph-embeddings, graph-learning, dgl, link-prediction, distributed-training, transe, rotate, linux, gpu

## Member repositories
- awslabs/dgl-ke (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:24.340917+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:45:15.252873+00:00, confidence not recorded.
  - readme: https://github.com/awslabs/dgl-ke (fetched 2026-08-28T04:04:24.340917+00:00, sha b72988f61135)
- Data as of 2026-08-30T08:39:29.467469+00:00.