iGCL Project Source Code

File Structure

~~ ➜ find . -maxdepth 2 -type d . ./Graph_Classification ./Node_Classification_Citation ./Node_Classification_Citation/citeseer ./Node_Classification_Citation/cora_pubmed ./Node_Classification_Amazon ~~

Requirements

torch_geometric
omegaconf
warmup_scheduler

Graph Classification

IMDB-BINARY

~~ cd ./Graph_Classification; python main_graph.py config/IMDB-BINARY.yml ~~

IMDB-MULTI

~~ cd ./Graph_Classification; python main_graph.py config/IMDB-MULTI.yml ~~

IMDB-MUTAG

~~ cd ./Graph_Classification; python main_graph.py config/MUTAG.yml ~~

NCI1

~~ cd ./Graph_Classification; python main_graph.py config/NCI1.yml ~~

COLLAB

~~ cd ./Graph_Classification; python main_graph.py config/COLLAB.yml ~~

PROTEINS

~~ cd ./Graph_Classification; python main_graph.py config/PROTEINS.yml ~~

Node Classification

Amazon Photos

~~ cd ./Node_Classification_Amazon python main.py --dataset Photo --lr 0.005 --epochs 600 --num_samples 3000 --vgae_lr 0.001 --lr_step_rate 0.85

lower gpu memory cost version

python main.py --dataset Photo --lr 0.005 --epochs 600 --num_samples 500 --vgae_lr 0.001 --lr_step_rate 0.85 ~~

Amazon Computers

~~ cd ./Node_Classification_Amazon python main.py --dataset Computers --lr 0.001 --epochs 600 --num_samples 3000 --vgae_lr 0.005 --lr_step_rate 0.85

lower gpu memory cost version

python main.py --dataset Computers --lr 0.001 --epochs 600 --num_samples 500 --vgae_lr 0.005 --lr_step_rate 0.85 ~~

Cora

~~ cd Node_Classification_Citation/cora_pubmed python reproduce_cora.py ~~

PubMed

~~ cd Node_Classification_Citation/cora_pubmed python reproduce_pubmed.py ~~

Citeseer

~~ cd Node_Classification_Citation/citeseer python reproduce_citeseer.py ~~