search

chennnM / GCNII

PyTorch implementation of "Simple and Deep Graph Convolutional Networks"
325 stars 72 forks source link

Some issues about the model #6

Open tanjia123456 opened 3 years ago

tanjia123456 commented 3 years ago

what's more, I read your code, where did you add the initial residual and identity mapping? Sorry, I’m just getting started, the question may be a bit silly

chennnM commented 3 years ago

what's more, I read your code, where did you add the initial residual and identity mapping? Sorry, I’m just getting started, the question may be a bit silly

Please refer to model.py (lines 30-36).

tanjia123456 commented 3 years ago

ok, thank you for your reply

tanjia123456 commented 3 years ago

hello when I run full-supervised. py，the accuracy is only 30%. what's reason of the output?

chennnM commented 3 years ago

hello when I run full-supervised. py，the accuracy is only 30%. what's reason of the output?

The default parameters in the code are incorrect. For Cora, please run python -u full-supervised.py --data cora --layer 64 --alpha 0.2 --weight_decay 1e-4 --variant

tanjia123456 commented 3 years ago

ok, thank you very much

zkcoder commented 1 year ago

hello,where is the data of Semi-supervised learning model?How about the unlabeled data? thank you very much