{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/graphgpt-graph-learning-with-generative-pre","title":"GraphGPT: Graph Learning with Generative Pre-trained Transformers","arxiv_id":"2401.00529","date":"2023-12-31","proceeding":null,"authors":["Qifang Zhao","Weidong Ren","Tianyu Li","Xiaoxiao Xu","Hong Liu"],"abstract":"We introduce \\textit{GraphGPT}, a novel model for Graph learning by self-supervised Generative Pre-training Transformers. Our model transforms each graph or sampled subgraph into a sequence of tokens representing the node, edge and attributes reversibly using the Eulerian path first. Then we feed the tokens into a standard transformer decoder and pre-train it with the next-token-prediction (NTP) task. Lastly, we fine-tune the GraphGPT model with the supervised tasks. This intuitive, yet effective model achieves superior or close results to the state-of-the-art methods for the graph-, edge- and node-level tasks on the large scale molecular dataset PCQM4Mv2, the protein-protein association dataset ogbl-ppa and the ogbn-proteins dataset from the Open Graph Benchmark (OGB). Furthermore, the generative pre-training enables us to train GraphGPT up to 400M+ parameters with consistently increasing performance, which is beyond the capability of GNNs and previous graph transformers. The source code and pre-trained checkpoints will be released soon\\footnote{\\url{https://github.com/alibaba/graph-gpt}} to pave the way for the graph foundation model research, and also to assist the scientific discovery in pharmaceutical, chemistry, material and bio-informatics domains, etc.","url_abs":"https://arxiv.org/abs/2401.00529v1","url_pdf":"https://arxiv.org/pdf/2401.00529v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"graphgpt-graph-learning-with-generative-pre","repo_url":"https://github.com/alibaba/graph-gpt","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"graph-learning","task_name":"Graph Learning"},{"task_slug":"scientific-discovery","task_name":"scientific discovery"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/link-property-prediction-on-ogbl-citation2","task":"Link Property Prediction","dataset":"ogbl-citation2","model":"GraphGPT(d1n30)","rank_in_archive_order":1,"of":23,"metrics":{"Ext. data":"No","Number of params":"133096832","Test MRR":"0.9305 ± 0.0020","Validation MRR":"0.9295 ± 0.0022"},"uses_additional_data":false},{"leaderboard":"/sota/link-property-prediction-on-ogbl-citation2","task":"Link Property Prediction","dataset":"ogbl-citation2","model":"GraphGPT(SMTP)","rank_in_archive_order":3,"of":23,"metrics":{"Ext. data":"No","Number of params":"46784128","Test MRR":"0.9055 ± 0.0016","Validation MRR":"0.9042 ± 0.0014"},"uses_additional_data":false},{"leaderboard":"/sota/link-property-prediction-on-ogbl-ppa","task":"Link Property Prediction","dataset":"ogbl-ppa","model":"GraphGPT(SMTP)","rank_in_archive_order":3,"of":26,"metrics":{"Ext. data":"No","Number of params":"145263360","Test Hits@100":"0.6876 ± 0.0067","Validation Hits@100":"0.7017 ± 0.0044"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2401.00529","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}