Leaderboards for Node Property Prediction
Check leaderboards for
- ogbn-products
- ogbn-proteins
- ogbn-arxiv
- ogbn-papers100M
- ogbn-mag
The bold method name indicates that the implementation is official (by the author of the original paper).
Package denotes the required package version for each dataset to be eligible for the leaderboard.
Leaderboard for ogbn-products
The classification accuracy on the test and validation sets. The higher, the better.
Package: >=1.1.1
| Rank | Method | Ext. data | Test Accuracy | Validation Accuracy | Contact | References | #Params | Hardware | Date |
|---|---|---|---|---|---|---|---|---|---|
| 1 | GLEM+GIANT+SAGN+SCR | Yes | 0.8737 ± 0.0006 | 0.9400 ± 0.0003 | Jianan Zhao (Mila & MSRA Team) | Paper, Code | 139,792,525 | Tesla V100 (32GB) | Oct 27, 2022 |
| 2 | LD+GIANT+SAGN+SCR | Yes | 0.8718 ± 0.0004 | 0.9399 ± 0.0002 | Zhihao Shi (MIRA Lab, USTC & CityBrain Lab, Alibaba Cloud) | Paper, Code | 110,636,896 | GeForce RTX 3090 (24GB GPU) | Sep 27, 2023 |
| 3 | **GraDBERT+GIANT & SAGN+SLE+CnS ** | Yes | 0.8692 ± 0.0007 | 0.9371 ± 0.0003 | Costas Mavromatis (UMN & AWS) | Paper, Code | 1,154,654 | GeForce RTX 3090 (24GB GPU) | Apr 20, 2023 |
| 4 | GIANT-XRT+R-SAGN+SCR+C&S | Yes | 0.8684 ± 0.0005 | 0.9365 ± 0.0003 | LeeXue (HIT Team) | Paper, Code | 1,154,142 | TITAN RTX (24GB GPU) | Sep 30, 2022 |
| 5 | GIANT-XRT+SAGN+SCR+C&S | Yes | 0.8680 ± 0.0007 | 0.9357 ± 0.0004 | Yufei He (CogDL Team) | Paper, Code | 1,154,654 | GeForce RTX 3090 24GB (GPU) | Jun 13, 2022 |
| 6 | GIANT-XRT+SAGN+MCR+C&S | Yes | 0.8673 ± 0.0008 | 0.9387 ± 0.0002 | Yufei He (CogDL Team) | Paper, Code | 1,154,654 | GeForce RTX™ 3090 24GB (GPU) | Dec 8, 2021 |
| 7 | GIANT-XRT+SAGN+SCR | Yes | 0.8667 ± 0.0009 | 0.9364 ± 0.0005 | Yufei He (CogDL Team) | Paper, Code | 1,154,654 | GeForce RTX™ 3090 24GB (GPU) | Jun 13, 2022 |
| 8 | GIANT-XRT+SAGN+MCR | Yes | 0.8651 ± 0.0009 | 0.9389 ± 0.0002 | Yufei He (CogDL Team) | Paper, Code | 1,154,654 | GeForce RTX™ 3090 24GB (GPU) | Dec 8, 2021 |
| 9 | LD+GAMLP | Yes | 0.8645 ± 0.0012 | 0.9415 ± 0.0003 | Zhihao Shi (MIRA Lab, USTC & CityBrain Lab, Alibaba Cloud) | Paper, Code | 144,331,677 | GeForce RTX 3090 (24GB GPU) | Sep 27, 2023 |
| 10 | GIANT-XRT+SAGN+SLE+C&S (use raw text) | Yes | 0.8643 ± 0.0020 | 0.9352 ± 0.0005 | Eli Chien (UIUC) | Paper, Code | 1,548,382 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 11 | GIANT-XRT+SAGN+SLE (use raw text) | Yes | 0.8622 ± 0.0022 | 0.9363 ± 0.0005 | Eli Chien (UIUC) | Paper, Code | 1,548,382 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 12 | GIANT-XRT+GAMLP+MCR | Yes | 0.8591 ± 0.0008 | 0.9402 ± 0.0004 | Yufei He (CogDL Team) | Paper, Code | 2,144,151 | GeForce RTX™ 3090 24GB (GPU) | Dec 8, 2021 |
| 13 | GAMLP+RLU+SCR+C&S | No | 0.8520 ± 0.0008 | 0.9304 ± 0.0005 | Yufei He (CogDL Team) | Paper, Code | 3,335,831 | GeForce RTX™ 3090 24GB (GPU) | Dec 8, 2021 |
| 14 | GAMLP+RLU+SCR | No | 0.8505 ± 0.0009 | 0.9292 ± 0.0005 | Yufei He (CogDL Team) | Paper, Code | 3,335,831 | GeForce RTX™ 3090 24GB (GPU) | Dec 8, 2021 |
| 15 | SAGN+SLE (4 stages)+C&S | No | 0.8485 ± 0.0010 | 0.9302 ± 0.0003 | Chuxiong Sun (CTRI) | Paper, Code | 2,179,678 | Tesla V100 (16GB GPU) | Sep 21, 2021 |
| 16 | SAGN+SLE (4 stages) | No | 0.8468 ± 0.0012 | 0.9309 ± 0.0007 | Chuxiong Sun (CTRI) | Paper, Code | 2,179,678 | Tesla V100 (16GB GPU) | Sep 21, 2021 |
| 17 | GAMLP+MCR | No | 0.8462 ± 0.0003 | 0.9319 ± 0.0003 | Yufei He (CogDL Team) | Paper, Code | 3,335,831 | GeForce RTX™ 3090 24GB (GPU) | Dec 8, 2021 |
| 18 | GAMLP+RLU | No | 0.8459 ± 0.0010 | 0.9324 ± 0.0005 | Wentao Zhang (PKU Tencent Joint Lab) | Paper, Code | 3,335,831 | Tesla V100 (32GB) | Aug 19, 2021 |
| 19 | Spec-MLP-Wide + C&S | No | 0.8451 ± 0.0006 | 0.9132 ± 0.0010 | Huixuan Chi (AML@ByteDance) | Paper, Code | 406,063 | Tesla V100 (32GB) | Jul 27, 2021 |
| 20 | SAGN+MCR | No | 0.8441 ± 0.0005 | 0.9325 ± 0.0004 | Yufei He (CogDL Team) | Paper, Code | 2,179,678 | GeForce RTX™ 3090 24GB (GPU) | Dec 8, 2021 |
| 21 | SAGN+SLE | No | 0.8428 ± 0.0014 | 0.9287 ± 0.0003 | Chuxiong Sun | Paper, Code | 2,179,678 | Tesla V100 (16GB GPU) | Apr 19, 2021 |
| 22 | MLP + C&S | No | 0.8418 ± 0.0007 | 0.9147 ± 0.0009 | Horace He (Cornell) | Paper, Code | 96,247 | GeForce RTX 2080 (11GB GPU) | Oct 27, 2020 |
| 23 | GIANT-XRT+GraphSAINT(use raw text) | Yes | 0.8415 ± 0.0022 | 0.9318 ± 0.0004 | Eli Chien (UIUC) | Paper, Code | 417,583 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 24 | Polynormer | No | 0.8382 ± 0.0011 | 0.9239 ± 0.0005 | Chenhui Deng | Paper, Code | 2,383,654 | NVIDIA RTX A6000 (48GB GPU) | Mar 20, 2024 |
| 25 | GAMLP | No | 0.8354 ± 0.0009 | 0.9312 ± 0.0003 | Wentao Zhang (PKU Tencent Joint Lab) | Paper, Code | 3,335,831 | Tesla V100 (32GB) | Aug 22, 2021 |
| 26 | AGDN | No | 0.8334 ± 0.0027 | 0.9229 ± 0.0010 | Chuxiong Sun | Paper, Code | 1,544,047 | Tesla V100 (16GB GPU) | Sep 2, 2022 |
| 27 | RevGNN-112 | No | 0.8307 ± 0.0030 | 0.9290 ± 0.0007 | Guohao Li - PyG Team | Paper, Code | 2,945,007 | NVIDIA Tesla V100 (32GB GPU) | May 27, 2022 |
| 28 | Linear + C&S | No | 0.8301 ± 0.0001 | 0.9134 ± 0.0001 | Horace He (Cornell) | Paper, Code | 10,763 | GeForce RTX 2080 (11GB GPU) | Oct 27, 2020 |
| 29 | UniMP | No | 0.8256 ± 0.0031 | 0.9308 ± 0.0017 | Yunsheng Shi (PGL team) | Paper, Code | 1,475,605 | Tesla V100 (32GB) | Sep 8, 2020 |
| 30 | Plain Linear + C&S | No | 0.8254 ± 0.0003 | 0.9103 ± 0.0001 | Horace He (Cornell) | Paper, Code | 4,747 | GeForce RTX 2080 (11GB GPU) | Oct 27, 2020 |
| 31 | DeeperGCN+FLAG | No | 0.8193 ± 0.0031 | 0.9221 ± 0.0037 | Kezhi Kong | Paper, Code | 253,743 | NVIDIA Tesla V100 (32GB GPU) | Oct 20, 2020 |
| 32 | GAT+FLAG | No | 0.8176 ± 0.0045 | 0.9251 ± 0.0006 | Kezhi Kong | Paper, Code | 751,574 | GeForce RTX 2080 Ti (11GB GPU) | Oct 20, 2020 |
| 33 | GraphSAGE + C&S + node2vec | No | 0.8154 ± 0.0050 | 0.9238 ± 0.0006 | HuiXuan Chi | Paper, Code | 103,983 | Tesla V100 (32GB) | Apr 6, 2021 |
| 34 | SAGN | No | 0.8120 ± 0.0007 | 0.9309 ± 0.0004 | Chuxiong Sun | Paper, Code | 2,233,391 | Tesla V100 (16GB GPU) | Apr 19, 2021 |
| 35 | PCAPass + XGBoost | No | 0.8115 ± 0.0002 | 0.9200 ± 0.0005 | Krzysztof Sadowski (Intel) | Paper, Code | 0 | Intel Xeon 8375C (CPU 128GB) | Feb 2, 2022 |
| 36 | DeeperGCN | No | 0.8098 ± 0.0020 | 0.9238 ± 0.0009 | Guohao Li - DeepGCNs.org | Paper, Code | 253,743 | NVIDIA Tesla V100 (32GB GPU) | Jun 28, 2020 |
| 36 | E2EG (use raw text) | Yes | 0.8098 ± 0.0040 | 0.9234 ± 0.0009 | Tu Anh Dinh | Paper, Code | 66,793,520 | GeForce GTX 1080 Ti (11GB GPU) | Aug 11, 2022 |
| 37 | GAT w/NS + C&S | No | 0.8092 ± 0.0037 | 0.9263 ± 0.0008 | HuiXuan Chi | Paper, Code | 753,622 | Tesla V100 (32GB) | Apr 4, 2021 |
| 38 | SIGN | No | 0.8052 ± 0.0016 | 0.9299 ± 0.0004 | Lingfan Yu (DGL Team) | Paper, Code | 3,483,703 | Tesla T4 (15 GB GPU) | Nov 5, 2020 |
| 39 | GIANT-XRT+MLP (use raw text) | Yes | 0.8049 ± 0.0028 | 0.9210 ± 0.0009 | Eli Chien (UIUC) | Paper, Code | 275,759 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 40 | GraphSAGE w/NS + C&S | No | 0.8041 ± 0.0022 | 0.9238 ± 0.0007 | HuiXuan Chi | Paper, Code | 207,919 | Tesla V100 (32GB) | Apr 5, 2021 |
| 41 | GraphSAINT-inductive | No | 0.8027 ± 0.0026 | Please tell us | Hanqing Zeng | Paper, Code | 331,661 | Tesla P100 (16GB GPU) | Jul 13, 2020 |
| 42 | ClusterGCN+residual+3 layers | No | 0.7971 ± 0.0042 | 0.9188 ± 0.0008 | Horace He (Cornell) | Paper, Code | 456,034 | GeForce RTX 2080 (11GB GPU) | Oct 6, 2020 |
| 43 | GAT with NeighborSampling | No | 0.7945 ± 0.0059 | Please tell us | Matthias Fey | Paper, Code | 751,574 | GeForce RTX 2080 (11GB GPU) | May 24, 2020 |
| 44 | GraphSAGE+FLAG | No | 0.7936 ± 0.0057 | 0.9205 ± 0.0007 | Kezhi Kong | Paper, Code | 206,895 | GeForce RTX 2080 Ti (11GB GPU) | Oct 20, 2020 |
| 45 | Cluster-GAT | No | 0.7923 ± 0.0078 | 0.8985 ± 0.0022 | Xiang Song | Paper, Code | 1,540,848 | EC2 P3.2xlarge (V100) | Aug 2, 2020 |
| 46 | GraphSAINT (SAGE aggr) | No | 0.7908 ± 0.0024 | 0.9162 ± 0.0008 | Matthias Fey – OGB team | Paper, Code | 206,895 | GeForce RTX 2080 (11GB GPU) | Jun 10, 2020 |
| 47 | ClusterGCN (SAGE aggr) | No | 0.7897 ± 0.0033 | 0.9212 ± 0.0009 | Matthias Fey – OGB team | Paper, Code | 206,895 | GeForce RTX 2080 (11GB GPU) | Jun 10, 2020 |
| 48 | NeighborSampling (SAGE aggr) | No | 0.7870 ± 0.0036 | 0.9170 ± 0.0009 | Matthias Fey – OGB team | Paper, Code | 206,895 | GeForce RTX 2080 (11GB GPU) | Jun 10, 2020 |
| 49 | Full-batch GraphSAGE | No | 0.7850 ± 0.0014 | 0.9224 ± 0.0007 | Matthias Fey – OGB team | Paper, Code | 206,895 | Quadro RTX 8000 (48GB GPU) | Jun 20, 2020 |
| 50 | GraphSAGE | No | 0.7829 ± 0.0016 | Please tell us | Quan Gan (DGL Team) | Paper, Code | Please tell us | Please tell us | May 12, 2020 |
| 51 | TCNN | No | 0.7606 ± 0.0037 | 0.8991 ± 0.0011 | Zheyi Qin (Colorado State University) | Paper, Code | 22,624 | NVIDIA A100 | Apr 5, 2024 |
| 52 | Full-batch GCN | No | 0.7564 ± 0.0021 | 0.9200 ± 0.0003 | Matthias Fey – OGB team | Paper, Code | 103,727 | Quadro RTX 8000 (48GB GPU) | Jun 20, 2020 |
| 53 | Label Propagation | No | 0.7434 ± 0.0000 | 0.9091 ± 0.0000 | Horace He (Cornell) | Paper, Code | 0 | GeForce RTX 2080 (11GB GPU) | Oct 3, 2020 |
| 54 | GraphZoom (Node2vec) | No | 0.7406 ± 0.0026 | 0.9066 ± 0.0011 | Xiuyu Li - GraphZoom | Paper, Code | 120,251,183 | NVIDIA TITAN RTX (24GB GPU) | Oct 6, 2020 |
| 55 | Node2vec | No | 0.7249 ± 0.0010 | 0.9032 ± 0.0006 | Matthias Fey – OGB team | Paper, Code | 313,612,207 | GeForce RTX 2080 (11GB GPU) | Jun 10, 2020 |
| 56 | CoLinkDistMLP | No | 0.6259 ± 0.0010 | 0.7721 ± 0.0015 | Yi Luo (UESTC) | Paper, Code | 115,806 | Geforce GTX 1080 Ti (11GB GPU) | Jun 17, 2021 |
| 57 | MLP+FLAG | No | 0.6241 ± 0.0016 | 0.7688 ± 0.0014 | Kezhi Kong | Paper, Code | 103,727 | GeForce RTX 2080 Ti (11GB GPU) | Nov 17, 2020 |
| 58 | MLP | No | 0.6106 ± 0.0008 | 0.7554 ± 0.0014 | Matthias Fey – OGB team | Paper, Code | 103,727 | GeForce RTX 2080 (11GB GPU) | Jun 10, 2020 |
Leaderboard for ogbn-proteins
The ROC-AUC score on the test and validation sets. The higher, the better.
Package: >=1.1.1
| Rank | Method | Ext. data | Test ROC-AUC | Validation ROC-AUC | Contact | References | #Params | Hardware | Date |
|---|---|---|---|---|---|---|---|---|---|
| 1 | LD+GAT | Yes | 0.8942 ± 0.0007 | 0.9527 ± 0.0007 | Zhihao Shi (MIRA Lab, USTC & CityBrain Lab, Alibaba Cloud) | Paper, Code | 664,233,700 | GeForce RTX 3090 (24GB GPU) | Sep 27, 2023 |
| 2 | GIPA(Wide&Deep) | No | 0.8917 ± 0.0007 | 0.9472 ± 0.0020 | Houyi Li | Paper, Code | 17,438,716 | Tesla V100-SXM2(32G) | Jan 19, 2023 |
| 3 | AGDN | No | 0.8865 ± 0.0013 | 0.9418 ± 0.0005 | Chuxiong Sun | Paper, Code | 8,605,486 | Tesla V100 (16GB GPU) | Sep 2, 2022 |
| 4 | RevGNN-Wide | No | 0.8824 ± 0.0015 | 0.9450 ± 0.0008 | Guohao Li - DeepGCNs.org | Paper, Code | 68,471,608 | NVIDIA RTX 6000 (48G) | Jun 16, 2021 |
| 5 | GAT+BOT+NGNN | No | 0.8809 ± 0.0016 | 0.9375 ± 0.0019 | Xiang song (DGL team) | Paper, Code | 11,740,552 | Tesla V100 (32GB) | Jan 23, 2022 |
| 6 | RevGNN-Deep | No | 0.8774 ± 0.0013 | 0.9326 ± 0.0006 | Guohao Li - DeepGCNs.org | Paper, Code | 20,031,384 | NVIDIA RTX 6000 (48G) | Jun 16, 2021 |
| 7 | GAT+BoT | No | 0.8765 ± 0.0008 | 0.9280 ± 0.0008 | Yangkun Wang (DGL Team) | Paper, Code | 2,484,192 | Tesla A100 (40GB GPU) | Jun 16, 2021 |
| 8 | GAT + labels + node2vec | No | 0.8711 ± 0.0007 | 0.9217 ± 0.0011 | Huixuan Chi | Paper, Code | 6,360,470 | Tesla V100 (32GB) | Jun 7, 2021 |
| 9 | GIPA | No | 0.8700 ± 0.0010 | 0.9187 ± 0.0003 | Qinkai Zheng (GeaLearn Team) | Paper, Code | 4,831,056 | GeForce Titan RTX (24GB GPU) | May 13, 2021 |
| 10 | UniMP+CrossEdgeFeat | No | 0.8691 ± 0.0018 | 0.9258 ± 0.0009 | Yelrose (PGL Team) | Paper, Code | 1,959,984 | Tesla V100 (32GB) | Nov 24, 2020 |
| 11 | GAT+EdgeFeatureAtt | No | 0.8682 ± 0.0021 | 0.9194 ± 0.0003 | Yangkun Wang (DGL Team) | Paper, Code | 2,475,232 | p3.8xlarge (15GB GPU) | Nov 6, 2020 |
| 12 | UniMP | No | 0.8642 ± 0.0008 | 0.9175 ± 0.0006 | Yunsheng Shi (PGL team) | Paper, Code | 1,909,104 | Tesla V100 (32GB) | Sep 8, 2020 |
| 13 | DeeperGCN+FLAG | No | 0.8596 ± 0.0027 | 0.9132 ± 0.0022 | Kezhi Kong | Paper, Code | 2,374,568 | GeForce RTX 2080 Ti (11GB GPU) | Oct 20, 2020 |
| 14 | DeeperGCN | No | 0.8580 ± 0.0017 | 0.9106 ± 0.0016 | Guohao Li - DeepGCNs.org | Paper, Code | 2,374,568 | NVIDIA Tesla V100 (32GB GPU) | Jun 16, 2020 |
| 15 | DeepGCN | No | 0.8496 ± 0.0028 | 0.8971 ± 0.0011 | Guohao Li - DeepGCNs.org | Paper, Code | 2,374,456 | NVIDIA Tesla V100 (32GB GPU) | Jun 20, 2020 |
| 16 | MWE-DGCN | No | 0.8436 ± 0.0065 | 0.8973 ± 0.0057 | Zhengdao Chen | Paper, Code | 538,544 | NVIDIA Tesla V100 (16GB GPU) | Jul 20, 2020 |
| 17 | GEN + FLAG + node2vec | No | 0.8251 ± 0.0043 | 0.8656 ± 0.0037 | HuiXuan Chi | Paper, Code | 487,436 | Tesla V100 (32GB) | Apr 15, 2021 |
| 18 | DVCNN | No | 0.7916 ± 0.0086 | 0.8256 ± 0.0057 | Zheyi Qin (CSU) | Paper, Code | 90,608 | NVIDIA A100 | Apr 5, 2024 |
| 19 | GeniePath-BS | No | 0.7825 ± 0.0035 | Please tell us | Zhengwei WU (AGL Team) | Paper, Code | 316,754 | Intel Xeon E5-2682 v4 (512GB CPU) | Jun 10, 2020 |
| 20 | GaAN | No | 0.7803 ± 0.0073 | Please tell us | Wenjin Wang (PGL Team) | Paper, Code | Please tell us | Please tell us | May 26, 2020 |
| 21 | GraphSAGE | No | 0.7768 ± 0.0020 | 0.8334 ± 0.0013 | Matthias Fey – OGB team | Paper, Code | 193,136 | GeForce RTX 2080 (11GB GPU) | May 1, 2020 |
| 22 | GCN | No | 0.7251 ± 0.0035 | 0.7921 ± 0.0018 | Matthias Fey – OGB team | Paper, Code | 96,880 | GeForce RTX 2080 (11GB GPU) | Jul 17, 2020 |
| 23 | MLP | No | 0.7204 ± 0.0048 | 0.7706 ± 0.0014 | Matthias Fey – OGB team | Paper, Code | 96,880 | GeForce RTX 2080 (11GB GPU) | May 1, 2020 |
| 24 | Node2vec | No | 0.6881 ± 0.0065 | 0.7007 ± 0.0053 | Matthias Fey – OGB team | Paper, Code | 17,094,000 | GeForce RTX 2080 (11GB GPU) | May 1, 2020 |
Leaderboard for ogbn-arxiv
The classification accuracy on the test and validation sets. The higher, the better.
Package: >=1.1.1
| Rank | Method | Ext. data | Test Accuracy | Validation Accuracy | Contact | References | #Params | Hardware | Date |
|---|---|---|---|---|---|---|---|---|---|
| 1 | SimTeG+TAPE+RevGAT | Yes | 0.7803 ± 0.0007 | 0.7846 ± 0.0004 | Keyu Duan | Paper, Code | 1,386,219,488 | 4 * A100-XMS4 (40GB GPU) | Aug 7, 2023 |
| 2 | TAPE+RevGAT | Yes | 0.7750 ± 0.0012 | 0.7785 ± 0.0016 | Xiaoxin He (NUS) | Paper, Code | 280,283,296 | 4 NVIDIA RTX A5000 24GB GPUs | May 31, 2023 |
| 3 | SimTeG+TAPE+GraphSAGE | Yes | 0.7748 ± 0.0011 | 0.7789 ± 0.0008 | Keyu Duan | Paper, Code | 1,381,593,403 | 4 * A100-XMS4 (40GB GPU) | Aug 7, 2023 |
| 4 | LD+REVGAT | Yes | 0.7726 ± 0.0017 | 0.7762 ± 0.0008 | Zhihao Shi (MIRA Lab, USTC & CityBrain Lab, Alibaba Cloud) | Paper, Code | 140,438,868 | GeForce RTX 3090 (24GB GPU) | Sep 27, 2023 |
| 5 | GraDBERT & RevGAT+KD | Yes | 0.7721 ± 0.0031 | 0.7757 ± 0.0009 | Costas Mavromatis (UMN & AWS) | Paper, Code | 1,304,912 | GeForce RTX 3090 (24GB GPU) | Apr 20, 2023 |
| 6 | GLEM+RevGAT | Yes | 0.7694 ± 0.0025 | 0.7746 ± 0.0018 | Jianan Zhao (Mila & MSRA Team) | Paper, Code | 140,469,624 | Tesla V100 (32GB) | Oct 27, 2022 |
| 7 | GIANT-XRT+AGDN+BoT+self-KD | Yes | 0.7637 ± 0.0011 | 0.7719 ± 0.0008 | Chuxiong Sun | Paper, Code | 1,309,760 | Tesla V100 (16GB GPU) | Sep 2, 2022 |
| 8 | GIANT-XRT+RevGAT+KD+DCN | Yes | 0.7636 ± 0.0013 | 0.7699 ± 0.0002 | Xiaojun Guo(xjguo) | Paper, Code | 1,304,912 | GeForce GTX 1080 Ti(12GB GPU) | Apr 24, 2023 |
| 9 | GIANT-XRT+R-RevGAT+KD | Yes | 0.7635 ± 0.0006 | 0.7692 ± 0.0010 | LeeXue (HIT Team) | Paper, Code | 1,500,712 | TITAN RTX (24GB GPU) | Sep 30, 2022 |
| 10 | GIANT-XRT+DRGAT+KD | Yes | 0.7633 ± 0.0008 | 0.7725 ± 0.0006 | anonymous_zhang(anonymous) | Paper, Code | 2,685,527 | Tesla P100-PCIE-16GB | Jan 14, 2022 |
| 11 | GIANT-XRT+AGDN+BoT | Yes | 0.7618 ± 0.0016 | 0.7724 ± 0.0006 | Chuxiong Sun | Paper, Code | 1,309,760 | Tesla V100 (16GB GPU) | Sep 2, 2022 |
| 12 | GIANT-XRT+RevGAT+KD (use raw text) | Yes | 0.7615 ± 0.0010 | 0.7716 ± 0.0009 | Eli Chien (UIUC) | Paper, Code | 1,304,912 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 13 | GIANT-XRT+DRGAT | No | 0.7611 ± 0.0009 | 0.7716 ± 0.0008 | anonymous_zhang(anonymous) | Paper, Code | 2,685,527 | Tesla P100-PCIE-16GB | Jan 17, 2022 |
| 14 | GIANT-XRT+RevGAT (use raw text) | Yes | 0.7590 ± 0.0019 | 0.7701 ± 0.0009 | Eli Chien (UIUC) | Paper, Code | 1,304,912 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 15 | LGGNN+LabelReuse+C&S | No | 0.7570 ± 0.0018 | 0.7687 ± 0.0005 | Shichao Ma(Topo@OppoResearch) | Paper, Code | 1,161,640 | Tesla V100 (32GB) | Nov 3, 2022 |
| 15 | GIANT-XRT+LGGNN+LabelReuse+C&S | Yes | 0.7570 ± 0.0018 | 0.7687 ± 0.0005 | Shichao Ma(Topo@OppoResearch) | Paper, Code | 1,161,640 | Tesla V100 (32GB) | Nov 3, 2022 |
| 16 | SciBERT & EHGCN (use MAG data) | Yes | 0.7461 ± 0.0006 | 0.7586 ± 0.0012 | Khang Ly (UvA, Elsevier) | Paper, Code | 621,944 | Tesla T4 (16GB GPU) | Jul 17, 2023 |
| 17 | GIANT-XRT+GraphSAGE (use raw text) | Yes | 0.7435 ± 0.0014 | 0.7595 ± 0.0011 | Eli Chien (UIUC) | Paper, Code | 546,344 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 18 | AGDN+BoT+self-KD+C&S | No | 0.7431 ± 0.0014 | 0.7518 ± 0.0009 | Chuxiong Sun | Paper, Code | 1,513,294 | Tesla V100 (16GB GPU) | Jul 22, 2021 |
| 19 | AGDN+BoT+self-KD | No | 0.7428 ± 0.0017 | 0.7526 ± 0.0001 | Chuxiong Sun | Paper, Code | 1,513,294 | Tesla V100 (16GB GPU) | Jul 22, 2021 |
| 20 | RevGAT+N.Adj+LabelReuse+SelfKD | No | 0.7426 ± 0.0017 | 0.7497 ± 0.0008 | Guohao Li - DeepGCNs.org | Paper, Code | 2,098,256 | NVIDIA Tesla V100 (32GB GPU) | Jun 21, 2021 |
| 21 | GAT-node2vec + BoT + self-KD | No | 0.7420 ± 0.0004 | 0.7482 ± 0.0015 | Huixuan Chi | Paper, Code | 1,700,432 | Tesla V100 (32GB) | Jun 28, 2021 |
| 22 | DRGAT | No | 0.7416 ± 0.0007 | 0.7534 ± 0.0002 | anonymous_zhang(anonymous) | Paper, Code | 2,685,527 | Tesla P100-PCIE-16GB | Aug 5, 2022 |
| 22 | GAT+label reuse+self KD | No | 0.7416 ± 0.0008 | 0.7514 ± 0.0004 | Shunli Ren(CMIC@SJTU) | Paper, Code | 1,441,580 | GeForce RTX 1080Ti (11GB GPU) | Dec 15, 2020 |
| 23 | GIANT+XRT+GATv2 | Yes | 0.7415 ± 0.0005 | 0.7527 ± 0.0008 | Leslie(JNU) | Paper, Code | 207,520 | Tesla T4(16G) | Jun 8, 2022 |
| 24 | AGDN+BoT | No | 0.7410 ± 0.0015 | 0.7522 ± 0.0007 | Chuxiong Sun | Paper, Code | 1,513,294 | Tesla V100 (16GB GPU) | Jul 22, 2021 |
| 25 | GAT-node2vec + BoT | No | 0.7405 ± 0.0004 | 0.7482 ± 0.0015 | Huixuan Chi | Paper, Code | 1,700,432 | Tesla V100 (32GB) | Jun 28, 2021 |
| 26 | RevGAT+NormAdj+LabelReuse | No | 0.7402 ± 0.0018 | 0.7501 ± 0.0010 | Guohao Li - DeepGCNs.org | Paper, Code | 2,098,256 | NVIDIA Tesla V100 (32GB GPU) | Jun 21, 2021 |
| 27 | GAT+label+reuse+topo loss | No | 0.7399 ± 0.0012 | 0.7513 ± 0.0009 | Mengyang Niu (DAMO DI) | Paper, Code | 1,441,580 | Tesla V100 (16GB) | Dec 10, 2020 |
| 28 | AGDN (GAT-HA+3_heads+labels) | No | 0.7398 ± 0.0009 | 0.7519 ± 0.0009 | Chuxiong Sun | Paper, Code | 1,508,555 | Tesla V100 (32GB GPU) | Jan 3, 2021 |
| 29 | UniMP_v2 | No | 0.7397 ± 0.0015 | 0.7506 ± 0.0009 | Weiyue Su (PGL Team) | Paper, Code | 687,377 | Tesla V100 (32GB) | Nov 24, 2020 |
| 30 | GAT(norm.adj.)+label reuse+C&S | No | 0.7395 ± 0.0012 | 0.7519 ± 0.0008 | Yangkun Wang (DGL Team) | Paper, Code | 1,441,580 | p3.8xlarge (15GB GPU) | Nov 24, 2020 |
| 31 | GAT+norm. adj.+label reuse | No | 0.7391 ± 0.0012 | 0.7516 ± 0.0008 | Yangkun Wang (DGL Team) | Paper, Code | 1,441,580 | p3.8xlarge (15GB GPU) | Nov 11, 2020 |
| 32 | GAT + C&S | No | 0.7386 ± 0.0014 | 0.7484 ± 0.0007 | Horace He (Cornell) | Paper, Code | 1,567,000 | GeForce RTX 2080 (11GB GPU) | Oct 27, 2020 |
| 33 | UniMP_large | No | 0.7379 ± 0.0014 | 0.7475 ± 0.0008 | Yunsheng Shi (PGL team) | Paper, Code | 1,162,515 | Tesla V100 (32GB) | Sep 25, 2020 |
| 34 | AGDN (GAT-HA+3_heads) | No | 0.7375 ± 0.0021 | 0.7483 ± 0.0009 | Chuxiong Sun | Paper, Code | 1,447,115 | Tesla V100 (32GB GPU) | Jan 3, 2021 |
| 35 | GAT+FLAG | No | 0.7371 ± 0.0013 | 0.7496 ± 0.0010 | Kezhi Kong | Paper, Code | 1,628,440 | GeForce RTX 2080 Ti (11GB GPU) | Oct 20, 2020 |
| 35 | LEGNN + AS-Train | No | 0.7371 ± 0.0011 | 0.7494 ± 0.0008 | Le Yu (Beihang University) | Paper, Code | 5,374,120 | NVIDIA Tesla T4 (15 GB) | May 31, 2022 |
| 36 | GAT+norm. adj.+labels | No | 0.7366 ± 0.0011 | 0.7508 ± 0.0009 | Yangkun Wang (DGL Team) | Paper, Code | 1,441,580 | p3.8xlarge (15GB GPU) | Oct 29, 2020 |
| 37 | GAT+norm.adj.+labels | No | 0.7365 ± 0.0011 | 0.7504 ± 0.0006 | Yangkun Wang (DGL Team) | Paper, Code | 1,628,440 | p3.8xlarge (15GB GPU) | Sep 17, 2020 |
| 38 | E2EG (use raw text) | Yes | 0.7362 ± 0.0014 | 0.7487 ± 0.0011 | Tu Anh Dinh | Paper, Code | 83,724,841 | GeForce GTX 1080 Ti (11GB GPU) | Aug 11, 2022 |
| 39 | Polynormer | No | 0.7346 ± 0.0016 | 0.7459 ± 0.0010 | Chenhui Deng (Cornell) | Paper, Code | 1,806,160 | NVIDIA RTX A6000 (48GB GPU) | Mar 20, 2024 |
| 40 | LEGNN | No | 0.7337 ± 0.0007 | 0.7480 ± 0.0009 | Le Yu (Beihang University) | Paper, Code | 5,374,120 | NVIDIA Tesla T4 (15 GB) | May 31, 2022 |
| 41 | GCN_res + C&S_v2 | No | 0.7313 ± 0.0017 | 0.7445 ± 0.0011 | HuiXuan Chi | Paper, Code | 155,824 | Tesla V100 (32GB) | Mar 26, 2021 |
| 42 | MLP + C&S | No | 0.7312 ± 0.0012 | 0.7391 ± 0.0015 | Horace He (Cornell) | Paper, Code | 175,656 | GeForce RTX 2080 (11GB GPU) | Oct 27, 2020 |
| 43 | UniMP | No | 0.7311 ± 0.0020 | 0.7450 ± 0.0005 | Yunsheng Shi (PGL team) | Paper, Code | 473,489 | Tesla V100 (32GB) | Sep 8, 2020 |
| 44 | GCN+linear+labels | No | 0.7306 ± 0.0024 | 0.7442 ± 0.0012 | Yangkun Wang (DGL Team) | Paper, Code | 238,632 | g4dn.12xlarge, T4 (15GB GPU) | Sep 5, 2020 |
| 44 | GIANT-XRT+MLP (use raw text) | Yes | 0.7306 ± 0.0011 | 0.7432 ± 0.0009 | Eli Chien (UIUC) | Paper, Code | 273,960 | Tesla T4 (16GB GPU) | Nov 8, 2021 |
| 45 | GTAN | No | 0.7297 ± 0.0017 | 0.7384 ± 0.0007 | Chaofan Wang | Paper, Code | 39,208 | RTX 3060 (12GB GPU) | May 18, 2022 |
| 45 | GCN_res + C&S | No | 0.7297 ± 0.0022 | 0.7423 ± 0.0014 | HuiXuan Chi | Paper, Code | 155,824 | Tesla T4（16GB） | Mar 24, 2021 |
| 46 | GCN+residual+6 layers | No | 0.7286 ± 0.0016 | 0.7382 ± 0.0007 | Horace He (Cornell) | Paper, Code | 122,542 | GeForce RTX 2080 (11GB GPU) | Oct 6, 2020 |
| 47 | GCN+residual+node2vec | No | 0.7278 ± 0.0013 | 0.7414 ± 0.0008 | Horace He (Cornell) | Paper, Code | 21,885,098 | GeForce RTX 2080 (11GB GPU) | Oct 3, 2020 |
| 48 | GCN_res + 8 layers + FLAG | No | 0.7276 ± 0.0024 | 0.7389 ± 0.0012 | Huixuan Chi | Paper, Code | 155,824 | Tesla T4（16GB） | Feb 23, 2021 |
| 49 | GCNII | No | 0.7274 ± 0.0016 | Please tell us | Ming Chen | Paper, Code | 2,148,648 | Quadro RTX 8000 (48GB GPU) | Jul 7, 2020 |
| 50 | GCN_res + 8 layers | No | 0.7262 ± 0.0037 | 0.7369 ± 0.0021 | Huixuan Chi | Paper, Code | 155,824 | Tesla T4（16GB） | Feb 20, 2021 |
| 51 | GTCN | No | 0.7225 ± 0.0017 | 0.7320 ± 0.0005 | Chaofan Wang | Paper, Code | 109,096 | RTX 3060 (12GB GPU) | May 18, 2022 |
| 52 | Linear + C&S | No | 0.7222 ± 0.0002 | 0.7368 ± 0.0004 | Horace He | Paper, Code | 15,400 | GeForce RTX 2080 (11GB GPU) | Oct 27, 2020 |
| 53 | JKNet (GCN-based) | No | 0.7219 ± 0.0021 | 0.7335 ± 0.0007 | Weiran Huang | Paper, Code | 89,000 | Tesla T4 | Aug 26, 2020 |
| 53 | EGC-S (100k) | No | 0.7219 ± 0.0016 | 0.7338 ± 0.0022 | Shyam Tailor | Paper, Code | 100,648 | GTX1080Ti/RTX2080Ti | Apr 6, 2021 |
| 53 | GraphSAGE+FLAG | No | 0.7219 ± 0.0021 | 0.7349 ± 0.0009 | Kezhi Kong | Paper, Code | 218,664 | GeForce RTX 2080 Ti (11GB GPU) | Oct 20, 2020 |
| 54 | DeeperGCN+FLAG | No | 0.7214 ± 0.0019 | 0.7311 ± 0.0009 | Kezhi Kong | Paper, Code | 491,176 | NVIDIA Tesla V100 (32GB GPU) | Oct 20, 2020 |
| 55 | DAGNN | No | 0.7209 ± 0.0025 | 0.7290 ± 0.0011 | Meng Liu - DIVE@TAMU | Paper, Code | 43,857 | GeForce RTX 2080 Ti (11GB GPU) | Aug 19, 2020 |
| 56 | GCN+FLAG | No | 0.7204 ± 0.0020 | 0.7330 ± 0.0010 | Kezhi Kong | Paper, Code | 142,888 | GeForce RTX 2080 Ti (11GB GPU) | Oct 20, 2020 |
| 57 | GaAN | No | 0.7197 ± 0.0024 | Please tell us | Hui Zhong (PGL Team) | Paper, Code | 1,471,506 | NVIDIA Tesla V100 (16GB GPU) | Jun 16, 2020 |
| 58 | EGC-M (100k) | No | 0.7196 ± 0.0023 | 0.7334 ± 0.0013 | Shyam Tailor | Paper, Code | 99,464 | GTX1080Ti/RTX2080Ti | Apr 6, 2021 |
| 59 | SIGN | No | 0.7195 ± 0.0011 | 0.7323 ± 0.0006 | Lingfan Yu (DGL Team) | Paper, Code | 3,566,128 | Tesla T4 (15GB) | Nov 5, 2020 |
| 60 | DeeperGCN | No | 0.7192 ± 0.0016 | 0.7262 ± 0.0014 | Guohao Li - DeepGCNs.org | Paper, Code | 491,176 | NVIDIA Tesla V100 (32GB GPU) | Jun 16, 2020 |
| 61 | PCAPass + XGBoost | No | 0.7187 ± 0.0003 | 0.7325 ± 0.0005 | Krzysztof Sadowski (Intel) | Paper, Code | 0 | Intel Xeon 8375C (CPU 128GB) | Feb 2, 2022 |
| 62 | GCN | No | 0.7174 ± 0.0029 | 0.7300 ± 0.0017 | Matthias Fey – OGB team | Paper, Code | 110,120 | GeForce RTX 2080 (11GB GPU) | May 1, 2020 |
| 63 | GraphSAGE | No | 0.7149 ± 0.0027 | 0.7277 ± 0.0016 | Matthias Fey – OGB team | Paper, Code | 218,664 | GeForce RTX 2080 (11GB GPU) | May 1, 2020 |
| 64 | Plain Linear + C&S | No | 0.7126 ± 0.0001 | 0.7300 ± 0.0001 | Horace He (Cornell) | Paper, Code | 5,160 | GeForce RTX 2080 (11GB GPU) | Oct 27, 2020 |
| 65 | GraphZoom (Node2vec) | No | 0.7118 ± 0.0018 | 0.7220 ± 0.0007 | Xiuyu Li - GraphZoom | Paper, Code | 8,963,624 | GeForce GTX 1080 Ti (11GB GPU) | Jul 2, 2020 |
| 66 | Node2vec | No | 0.7007 ± 0.0013 | 0.7129 ± 0.0013 | Matthias Fey – OGB team | Paper, Code | 21,818,792 | GeForce RTX 2080 (11GB GPU) | May 1, 2020 |
| 67 | Label Propagation | No | 0.6832 ± 0.0000 | 0.7014 ± 0.0000 | Horace He (Cornell) | Paper, Code | 0 | GeForce RTX 2080 (11GB GPU) | Oct 2, 2020 |
| 68 | CoLinkDistMLP | No | 0.5638 ± 0.0016 | 0.5807 ± 0.0011 | Yi Luo (UESTC) | Paper, Code | 120,912 | Geforce GTX 1080 Ti (11GB GPU) | Jun 17, 2021 |
| 69 | MLP+FLAG | No | 0.5602 ± 0.0019 | 0.5817 ± 0.0011 | Kezhi Kong | Paper, Code | 110,120 | GeForce RTX 2080 Ti (11GB GPU) | Oct 20, 2020 |
| 70 | MLP | No | 0.5550 ± 0.0023 | 0.5765 ± 0.0012 | Matthias Fey – OGB team | Paper, Code | 110,120 | GeForce RTX 2080 (11GB GPU) | May 1, 2020 |
Leaderboard for ogbn-papers100M
The classification accuracy on the test and validation sets. The higher, the better.
Package: >=1.2.0
| Rank | Method | Ext. data | Test Accuracy | Validation Accuracy | Contact | References | #Params | Hardware | Date |
|---|---|---|---|---|---|---|---|---|---|
| 1 | GLEM+GIANT+GAMLP | Yes | 0.7037 ± 0.0002 | 0.7354 ± 0.0001 | Jianan Zhao (Mila & MSRA Team) | Paper, Code | 154,775,375 | Tesla V100 (32GB) | Nov 9, 2022 |
| 2 | GIANT-XRT+GAMLP+RLU (use raw text) | Yes | 0.6967 ± 0.0005 | 0.7305 ± 0.0004 | Wei-Cheng Chang (Amazon) | Paper, Code | 21,551,631 | Tesla V100 (32GB GPU) | Nov 11, 2021 |
| 3 | GAMLP+RLU+SCR | No | 0.6842 ± 0.0015 | 0.7188 ± 0.0007 | Yufei He (CogDL Team) | Paper, Code | 67,560,875 | GeForce RTX 3090 24GB (GPU) | Jun 13, 2022 |
| 4 | SAGN+SLE (4 stages) | No | 0.6830 ± 0.0008 | 0.7163 ± 0.0007 | Chuxiong Sun (CTRI) | Paper, Code | 8,556,888 | Tesla V100 (16GB GPU) | Sep 21, 2021 |
| 5 | GAMLP+RLU | No | 0.6825 ± 0.0011 | 0.7159 ± 0.0005 | Wentao Zhang (PKU Tencent Joint Lab) | Paper, Code | 16,308,751 | Tesla V100 (32GB) | Aug 19, 2021 |
| 6 | GAMLP+SCR-m | No | 0.6816 ± 0.0012 | 0.7186 ± 0.0008 | Yufei He (CogDL Team) | Paper, Code | 67,560,875 | GeForce RTX 3090 24GB (GPU) | Jun 13, 2022 |
| 7 | GAMLP+SCR | No | 0.6814 ± 0.0008 | 0.7190 ± 0.0007 | Yufei He (CogDL Team) | Paper, Code | 67,560,875 | GeForce RTX 3090 24GB (GPU) | Jun 13, 2022 |
| 8 | FSGNN | No | 0.6807 ± 0.0006 | 0.7175 ± 0.0007 | Sunil Kumar Maurya (TokyoTech, AIST) | Paper, Code | 16,453,301 | NVIDIA V100 (16GB) | Sep 16, 2021 |
| 9 | SAGN+SLE | No | 0.6800 ± 0.0015 | 0.7131 ± 0.0010 | Chuxiong Sun | Paper, Code | 8,556,888 | Tesla V100 (16GB GPU) | Apr 19, 2021 |
| 10 | GAMLP | No | 0.6771 ± 0.0020 | 0.7117 ± 0.0014 | Wentao Zhang (PKU Tencent Joint Lab) | Paper, Code | 16,308,751 | Tesla V100 (32GB) | Aug 22, 2021 |
| 11 | TransformerConv | No | 0.6736 ± 0.0010 | 0.7172 ± 0.0005 | Xiaonan Song (NVIDIA SAE China team) | Paper, Code | 883,378 | NVIDIA DGX-2 (16*32GB GPUs) | Mar 4, 2021 |
| 12 | GraphSAGE_res_incep | No | 0.6706 ± 0.0017 | 0.7032 ± 0.0011 | Mengyang Niu (DAMO DI) | Paper, Code | 5,755,172 | Tesla V100 (16GB) | Feb 28, 2021 |
| 13 | SAGN | No | 0.6675 ± 0.0084 | 0.7034 ± 0.0099 | Chuxiong Sun | Paper, Code | 6,098,092 | Tesla V100 (16GB GPU) | Apr 19, 2021 |
| 14 | SIGN-XL | No | 0.6606 ± 0.0019 | 0.6984 ± 0.0006 | Fabrizio Frasca | Paper, Code | 7,180,460 | NVIDIA K80 GPU (12GB GPU) | Nov 4, 2020 |
| 15 | PCAPass + LightGBM | No | 0.6591 ± 0.0003 | 0.6982 ± 0.0002 | Krzysztof Sadowski (Intel) | Paper, Code | 0 | Intel Xeon 6330 (CPU 1TB) | Feb 9, 2022 |
| 16 | SIGN | No | 0.6568 ± 0.0006 | 0.6932 ± 0.0006 | Emanuele Rossi | Paper, Code | 1,008,812 | NVIDIA K80 GPU (12GB GPU) | Nov 4, 2020 |
| 17 | SGC | No | 0.6329 ± 0.0019 | 0.6648 ± 0.0020 | Weihua Hu – OGB team | Paper, Code | 144,044 | Xeon E7-8890x (1.5TB CPU) | Jun 10, 2020 |
| 18 | Node2vec | No | 0.5560 ± 0.0023 | 0.5807 ± 0.0028 | Weihua Hu – OGB team | Paper, Code | 14,215,818,412 | Xeon E7-8890x (1.5TB CPU) | Jun 26, 2020 |
| 19 | MLP | No | 0.4724 ± 0.0031 | 0.4960 ± 0.0029 | Weihua Hu – OGB team | Paper, Code | 144,044 | Xeon E7-8890x (1.5TB CPU) | Jun 10, 2020 |
Leaderboard for ogbn-mag
The classification accuracy on the test and validation sets. The higher, the better.
Package: >=1.2.1
| Rank | Method | Ext. data | Test Accuracy | Validation Accuracy | Contact | References | #Params | Hardware | Date |
|---|---|---|---|---|---|---|---|---|---|
| 1 | CLGNN | No | 0.7956 ± 0.0047 | 0.8021 ± 0.0020 | Zhen Hao Wong (E. Copi & EE Tsinghua) | Paper, Code | 7,720,368 | NVIDIA GeForce RTX 3090 | Feb 29, 2024 |
| 2 | LMSPS (w/o embs) | No | 0.5784 ± 0.0022 | 0.5951 ± 0.0007 | Chao Li (HUST-HCCS) | Paper, Code | 16,470,044 | Tesla V100 (16GB) | Feb 5, 2024 |
| 3 | RpHGNN+LP+CR (LINE embs) | No | 0.5773 ± 0.0012 | 0.5973 ± 0.0008 | Jun Hu (NUS) | Paper, Code | 7,720,368 | GeForce GTX 1080 Ti (11GB GPU) | Oct 24, 2023 |
| 4 | PSHGCN | No | 0.5752 ± 0.0011 | 0.5943 ± 0.0015 | Mingguo He (RUC & PGL Team) | Paper, Code | 4,852,434 | Tesla A100 (80 GB GPU) | Jun 8, 2023 |
| 5 | SeHGNN (ComplEx embs) | No | 0.5719 ± 0.0012 | 0.5917 ± 0.0009 | Xiaocheng Yang (ICT-GIMLab) | Paper, Code | 8,371,231 | NVIDIA Tesla T4 (15 GB) | Jul 7, 2022 |
| 6 | SeHGNN | No | 0.5671 ± 0.0014 | 0.5870 ± 0.0008 | Xiaocheng Yang (ICT-GIMLab) | Paper, Code | 8,371,231 | NVIDIA Tesla T4 (15 GB) | Jul 7, 2022 |
| 7 | NARS-GAMLP+RLU+SCR | No | 0.5631 ± 0.0021 | 0.5734 ± 0.0035 | Yufei He (CogDL Team) | Paper, Code | 6,734,882 | GeForce RTX 3090 24GB (GPU) | Jun 13, 2022 |
| 8 | NARS-GAMLP+RLU | No | 0.5590 ± 0.0027 | 0.5702 ± 0.0041 | Wentao Zhang (PKU Tencent Joint Lab) | Paper, Code | 6,734,882 | Tesla V100 (32GB) | Aug 19, 2021 |
| 9 | NARS-GAMLP+SCR-m | No | 0.5451 ± 0.0019 | 0.5590 ± 0.0028 | Yufei He (CogDL Team) | Paper, Code | 6,734,882 | GeForce RTX 3090 24GB (GPU) | Jun 13, 2022 |
| 10 | NARS_SAGN+SLE | No | 0.5440 ± 0.0015 | 0.5591 ± 0.0017 | Chuxiong Sun | Paper, Code | 3,846,330 | Tesla V100 (16GB GPU) | Jun 29, 2021 |
| 11 | NARS-GAMLP+SCR | No | 0.5432 ± 0.0018 | 0.5654 ± 0.0021 | Yufei He (CogDL Team) | Paper, Code | 6,734,882 | GeForce RTX 3090 24GB (GPU) | Jun 13, 2022 |
| 12 | NARS-GAMLP | No | 0.5396 ± 0.0018 | 0.5548 ± 0.0008 | Wentao Zhang (PKU Tencent Joint Lab) | Paper, Code | 6,734,882 | Tesla V100 (32GB) | Aug 22, 2021 |
| 13 | LEGNN + AS-Train | No | 0.5378 ± 0.0016 | 0.5528 ± 0.0013 | Le Yu (Beihang University) | Paper, Code | 5,147,997 | NVIDIA Tesla T4 (15 GB) | May 31, 2022 |
| 14 | LEGNN | No | 0.5276 ± 0.0014 | 0.5443 ± 0.0009 | Le Yu (Beihang University) | Paper, Code | 5,147,997 | NVIDIA Tesla T4 (15 GB) | May 31, 2022 |
| 15 | NARS | No | 0.5240 ± 0.0016 | 0.5372 ± 0.0009 | Lingfan Yu | Paper, Code | 4,130,149 | Tesla T4 (15GB) | Feb 18, 2021 |
| 16 | R-HGNN | No | 0.5204 ± 0.0026 | 0.5361 ± 0.0022 | Le Yu | Paper, Code | 5,638,053 | NVIDIA Tesla T4 (15 GB) | May 24, 2021 |
| 17 | R-GSN + metapath2vec | No | 0.5109 ± 0.0038 | 0.5295 ± 0.0042 | Huixuan Chi | Paper, Code | 309,777,252 | Tesla V100 (32GB) | Jun 30, 2021 |
| 18 | HGConv | No | 0.5045 ± 0.0017 | 0.5300 ± 0.0018 | Le Yu | Paper, Code | 2,850,405 | NVIDIA TITAN Xp (12GB) | Feb 14, 2021 |
| 19 | R-GSN | No | 0.5032 ± 0.0037 | 0.5182 ± 0.0041 | Xinliang Wu | Paper, Code | 154,373,028 | GeForce GTX 1080Ti | Jan 29, 2021 |
| 20 | HGT (TransE embs) | No | 0.4982 ± 0.0013 | 0.5124 ± 0.0046 | Lingfan Yu | Paper, Code | 26,877,657 | Tesla T4 (15GB) | Feb 17, 2021 |
| 21 | GraphSAINT + metapath2vec | No | 0.4966 ± 0.0022 | 0.5066 ± 0.0017 | HuiXuan Chi | Paper, Code | 309,764,724 | Tesla V100 (32GB) | Apr 9, 2021 |
| 22 | HGT (LADIES Sample) | No | 0.4927 ± 0.0061 | 0.4989 ± 0.0047 | Ziniu Hu | Paper, Code | 21,173,389 | Tesla K80 (12GB GPU) | Jan 26, 2021 |
| 23 | GraphSAINT (R-GCN aggr) | No | 0.4751 ± 0.0022 | 0.4837 ± 0.0026 | Matthias Fey – OGB team | Paper, Code | 154,366,772 | GeForce RTX 2080 (11GB GPU) | Jun 26, 2020 |
| 24 | R-GCN+FLAG | No | 0.4737 ± 0.0048 | 0.4835 ± 0.0036 | Kezhi Kong | Paper, Code | 154,366,772 | GeForce RTX 2080 Ti (11GB GPU) | Oct 21, 2020 |
| 25 | NeighborSampling (R-GCN aggr) | No | 0.4678 ± 0.0067 | 0.4761 ± 0.0068 | Matthias Fey – OGB team | Paper, Code | 154,366,772 | GeForce RTX 2080 (11GB GPU) | Jun 26, 2020 |
| 26 | SIGN | No | 0.4046 ± 0.0012 | 0.4068 ± 0.0010 | Lingfan Yu (DGL Team) | Paper, Code | 3,724,645 | Tesla T4 (15GB GPU) | Nov 5, 2020 |
| 27 | Full-batch R-GCN | No | 0.3977 ± 0.0046 | 0.4084 ± 0.0041 | Matthias Fey – OGB team | Paper, Code | 154,366,772 | Quadro RTX 8000 (48GB GPU) | Jun 26, 2020 |
| 28 | ClusterGCN (R-GCN aggr) | No | 0.3732 ± 0.0037 | 0.3840 ± 0.0031 | Matthias Fey – OGB team | Paper, Code | 154,366,772 | GeForce RTX 2080 (11GB GPU) | Jun 26, 2020 |
| 29 | MetaPath2vec | No | 0.3544 ± 0.0036 | 0.3506 ± 0.0017 | Matthias Fey – OGB team | Paper, Code | 94,479,069 | GeForce RTX 2080 (11GB GPU) | Jun 26, 2020 |
| 30 | CoLinkDistMLP | No | 0.2761 ± 0.0018 | 0.2646 ± 0.0013 | Yi Luo (UESTC) | Paper, Code | 278,202 | Geforce GTX 1080 Ti (11GB GPU) | Jun 17, 2021 |
| 31 | MLP | No | 0.2692 ± 0.0026 | 0.2626 ± 0.0016 | Matthias Fey – OGB team | Paper, Code | 188,509 | GeForce RTX 2080 (11GB GPU) | Jun 26, 2020 |