search

lightaime / deep_gcns_torch

Pytorch Repo for DeepGCNs (ICCV'2019 Oral, TPAMI'2021), DeeperGCN (arXiv'2020) and GNN1000(ICML'2021): https://www.deepgcns.org
MIT License
1.13k stars 155 forks source link

About ogbg-molhiv results #87

Closed dongZheX closed 3 years ago

dongZheX commented 3 years ago

Hello! (^_^)， thanks for the code. I used the following command to train on the molhiv dataset： python main.py --use_gpu --conv_encode_edge --num_layers 7 --dataset ogbg-molhiv --block res+ --gcn_aggr softmax --t 1.0 --learn_t --dropout 0.2 Why I can't reproduce the results in LeaderBoard？Is it something wrong with me?

My result is : VAL_ROCAUC | TEST_ROCAUC 0.7976±0.0179 | 0.7744 ± 0.1234

LeaderBoard result: VAL_ROCAUC | TEST_ROCAUC 0.8427 ± 0.0063 | 0.7858 ± 0.0117

lightaime commented 3 years ago

Hi @dongZheX, could you please check again your results or provide the logs files you have. Our results on the ogb leaderboard is: Val: 0.8427 ± 0.0063, and Test: 0.7858 ± 0.0117. The val ROCAUC you got seems not reasonable.

dongZheX commented 3 years ago

Hi @dongZheX, could you please check again your results or provide the logs files you have. Our results on the ogb leaderboard is: Val: 0.8427 ± 0.0063, and Test: 0.7858 ± 0.0117. The val ROCAUC you got seems not reasonable.

It's my fault. I use a new method to record the results but have something wrong in it.