# THUDM/CogDL

CogDL: A Comprehensive Library for Graph Deep Learning (WWW 2023)

Repository: https://github.com/THUDM/CogDL
Canonical: https://ross.abutalabs.com/products/cogdl
Homepage: https://cogdl.ai
Language: Python
License: MIT
License Family: permissive
Topics: graph-neural-networks, pytorch, graph-embedding, node-classification, graph-classification, link-prediction, leaderboard, gnn-model
Last push: 2024-02-01T21:24:47+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2725, "days_push": 944, "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 1821, forks 308 (observed 2026-08-28T04:05:41.000015+00:00)

## What it is
CogDL is a Python/PyTorch toolkit for graph deep learning that lets researchers train and compare GNN and graph embedding models on tasks like node classification, graph classification, and link prediction. It provides optimized operators, hyper-parameter search, and a model leaderboard, published at WWW 2023.

## Use cases
- train graph neural networks for node classification
- benchmark GNN models on graph datasets
- learn graph embeddings for link prediction
- run graph classification experiments with hyper-parameter search
- experiment with graph self-supervised learning like GraphMAE
- apply GNN models to custom graph datasets

## When to choose
- you need a unified PyTorch framework for GNN research and baselines
- you want reproducible benchmarks across many graph models and datasets
- you need efficient, memory-optimized GNN training on GPU

## When to avoid
- you need general deep learning beyond graphs
- you want a production graph ML service rather than a research toolkit
- you need a framework under very active development - releases have slowed

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, benchmarking, data-science
- domain: machine-learning, deep-learning, data-science
- platform: python, cross-platform
- tags: graph-neural-networks, gnn, graph-embedding, node-classification, graph-classification, link-prediction, pytorch, graph-representation-learning, self-supervised-learning, algorithms, gpu

## Member repositories
- THUDM/CogDL (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:41.000015+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 2026-08-30T03:19:56.660235+00:00, confidence not recorded.
  - readme: https://github.com/THUDM/CogDL (fetched 2026-08-28T04:05:41.000015+00:00, sha b06dad35afbb)
  - homepage: https://cogdl.ai (fetched 2026-08-29T10:59:09.186747+00:00, sha 433beb600ca3)
  - registry_pypi: https://pypi.org/pypi/cogdl/json (fetched 2026-08-29T10:59:09.190835+00:00, sha 4e838b2edd4b)
- Data as of 2026-08-30T08:39:29.467469+00:00.