# graphdeeplearning/graphtransformer

Graph Transformer Architecture. Source code for "A Generalization of Transformer Networks to Graphs", DLG-AAAI'21.

Repository: https://github.com/graphdeeplearning/graphtransformer
Canonical: https://ross.abutalabs.com/products/graphtransformer
Homepage: https://arxiv.org/abs/2012.09699
Language: Python
License: MIT
License Family: permissive
Topics: graph-transformer, transformers, transformer, attention, graph-neural-networks, graph-deep-learning, aaai, transformer-networks
Last push: 2021-07-27T07:32:23+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2085, "days_push": 1863, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1028, forks 150 (observed 2026-08-28T04:03:17.502622+00:00)

## What it is
Reference implementation of the Graph Transformer architecture from the paper 'A Generalization of Transformer Networks to Graphs' (DLG-AAAI'21). It adapts transformer attention to arbitrary graphs using Laplacian eigenvector positional encodings, batch normalization, and edge feature representations.

## Use cases
- apply transformer architecture to graph-structured data
- run graph neural network experiments with positional encodings
- reproduce results from the graph transformer paper
- benchmark graph transformer models on standard graph datasets
- model molecular graphs with bond-type edge features
- learn how attention mechanisms generalize to graphs
- compare graph transformer against standard GNN baselines

## When to choose
- you need a reference implementation of graph transformers with Laplacian positional encodings
- you want to reproduce or extend the DLG-AAAI'21 paper results
- your graph data has meaningful edge features like chemical bond types
- you are researching attention-based models on arbitrary graph topologies

## When to avoid
- you need a production-ready, actively maintained GNN library
- you want a general-purpose graph deep learning framework rather than a single paper's code
- you need transformers for sequences or NLP rather than graphs
- you require long-term support or frequent updates, since the repo has not seen releases since 2021

## Facets
- artifact type: library
- maturity: maintenance
- function: deep-learning, machine-learning, benchmarking
- domain: machine-learning, deep-learning
- platform: python, cross-platform
- tags: graph-neural-networks, graph-transformer, attention, positional-encoding, laplacian-eigenvectors, research-code, aaai-2021, pytorch, molecular-graphs, edge-features, algorithms, research, gpu, linux

## Member repositories
- graphdeeplearning/graphtransformer (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:17.502622+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 2026-08-30T07:07:46.040759+00:00, confidence not recorded.
  - readme: https://github.com/graphdeeplearning/graphtransformer (fetched 2026-08-28T04:03:17.502622+00:00, sha 7b0fcbdccd4f)
  - homepage: https://arxiv.org/abs/2012.09699 (fetched 2026-08-29T13:08:03.768419+00:00, sha c5ae59792025)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T13:08:03.778312+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T13:08:03.782959+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T13:08:03.784870+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T13:08:03.781036+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.