{"id":13444175,"url":"https://github.com/pfnet-research/chainer-graph-cnn","last_synced_at":"2025-04-13T10:27:08.255Z","repository":{"id":86583086,"uuid":"84155174","full_name":"pfnet-research/chainer-graph-cnn","owner":"pfnet-research","description":"Chainer implementation of 'Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering' (https://arxiv.org/abs/1606.09375)","archived":false,"fork":false,"pushed_at":"2017-12-28T08:25:39.000Z","size":55,"stargazers_count":68,"open_issues_count":3,"forks_count":17,"subscribers_count":12,"default_branch":"master","last_synced_at":"2025-03-27T01:51:26.484Z","etag":null,"topics":["chainer","graph-convolutional-networks"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/pfnet-research.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null}},"created_at":"2017-03-07T04:37:10.000Z","updated_at":"2025-01-18T12:24:19.000Z","dependencies_parsed_at":null,"dependency_job_id":"30b15a83-0c64-471a-a8d4-cf31db5cab09","html_url":"https://github.com/pfnet-research/chainer-graph-cnn","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/pfnet-research%2Fchainer-graph-cnn","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/pfnet-research%2Fchainer-graph-cnn/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/pfnet-research%2Fchainer-graph-cnn/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/pfnet-research%2Fchainer-graph-cnn/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/pfnet-research","download_url":"https://codeload.github.com/pfnet-research/chainer-graph-cnn/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248697393,"owners_count":21147320,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["chainer","graph-convolutional-networks"],"created_at":"2024-07-31T03:02:20.965Z","updated_at":"2025-04-13T10:27:08.201Z","avatar_url":"https://github.com/pfnet-research.png","language":"Python","funding_links":[],"categories":["Python","Preferred Networks Research"],"sub_categories":["Services using Chainer"],"readme":"# Chainer Graph CNN\n\nThis is a Chainer implementation of\n_Defferrard et al., \"Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering\", NIPS 2016._\n(https://arxiv.org/abs/1606.09375)\n\nDisclaimer: PFN provides no warranty or support for this implementation. Use it at your own risk.\nSee [license](LICENSE) for details.\n\nThis is not the original author's implementation. This implementation was based on [https://github.com/mdeff/cnn_graph](https://github.com/mdeff/cnn_graph).\n\nUsage\n-----\n```\n# Trains a GraphCNN on MNIST\n$ python tools/train.py -c configs/default.json -o results -e 100 -g 0\n```\n\nPrerequisites\n-------------\n```\npip install -r requirements.txt\n```\nThis implementation has been tested with the following versions.\n```\npython 2.7.6\nchainer (1.19.0)\nnose (1.3.7)\nnumpy (1.11.3)\nscikit-learn (0.18.1)\nscipy (0.18.1)\n```\nIt may work with other versions; not tested.\n\nResults\n-------\n```\nUsing ADAM alpha=1e-4\nepoch       iteration   main/loss   main/accuracy  validation/main/loss  validation/main/accuracy\n1           600         0.515395    0.854901       0.193552              0.9453\n2           1200        0.195267    0.942567       0.122769              0.9652\n3           1800        0.139023    0.95875        0.0955012             0.9726\n4           2400        0.110456    0.9676         0.0769727             0.9762\n5           3000        0.0932845   0.972033       0.0643796             0.9812\n6           3600        0.0811693   0.975149       0.0603944             0.9824\n7           4200        0.074127    0.978266       0.0556359             0.9831\n8           4800        0.0670138   0.980266       0.0509385             0.9839\n9           5400        0.0625065   0.980933       0.0496262             0.9839\n10          6000        0.0585658   0.982366       0.0493765             0.9838\n11          6600        0.0547269   0.983082       0.0444783             0.9859\n12          7200        0.050334    0.984582       0.0413585             0.9866\n13          7800        0.0493707   0.985032       0.0416611             0.9873\n14          8400        0.0459602   0.985999       0.0437013             0.9859\n15          9000        0.044378    0.986715       0.0406627             0.987\n16          9600        0.0430196   0.986815       0.0394637             0.9866\n17          10200       0.0404675   0.988182       0.0385143             0.9877\n18          10800       0.0398833   0.988265       0.0366019             0.989\n19          11400       0.0371923   0.988998       0.0348309             0.9875\n20          12000       0.0361765   0.989215       0.0402662             0.9858\n-- snip --\n100         60000       0.0157423   0.995832       0.0292472             0.9901\n```\n\n```\nUsing ADAM alpha=1e-3\nepoch       iteration   main/loss   main/accuracy  validation/main/loss  validation/main/accuracy\n1           600         0.225126    0.930017       0.0767015             0.9768\n2           1200        0.0977682   0.969899       0.0606019             0.9801\n3           1800        0.0770546   0.976016       0.0513997             0.9838\n4           2400        0.0666313   0.979532       0.0424098             0.9866\n5           3000        0.06334     0.980782       0.051125              0.9841\n6           3600        0.0578026   0.982532       0.0457874             0.985\n7           4200        0.0541042   0.983982       0.0405522             0.9875\n8           4800        0.0514735   0.984432       0.0443701             0.9867\n9           5400        0.0503822   0.984448       0.0557598             0.9812\n10          6000        0.0465654   0.985432       0.035589              0.9897\n11          6600        0.0455079   0.985932       0.03442               0.988\n12          7200        0.0425339   0.986882       0.038998              0.9868\n13          7800        0.0427513   0.986899       0.0395496             0.9873\n14          8400        0.0431217   0.986815       0.0372915             0.9877\n15          9000        0.0420674   0.987432       0.0401286             0.9864\n16          9600        0.0408353   0.987482       0.0404751             0.9876\n17          10200       0.0401931   0.987515       0.0372056             0.9879\n18          10800       0.0388781   0.988315       0.0389307             0.9889\n19          11400       0.0391798   0.988198       0.0406604             0.9872\n20          12000       0.0380889   0.988298       0.039208              0.9867\n-- snip --\n100         60000       0.0320832   0.990331       0.0345484             0.9887\n```\n\nLicense\n-------\nMIT License. Please see the LICENSE file for details.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fpfnet-research%2Fchainer-graph-cnn","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fpfnet-research%2Fchainer-graph-cnn","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fpfnet-research%2Fchainer-graph-cnn/lists"}