Skip to content

Latest commit

 

History

History
57 lines (42 loc) · 1.94 KB

README.md

File metadata and controls

57 lines (42 loc) · 1.94 KB
Raw

Dependencies

Then, you need to create a directory for recoreding finetuned results to avoid errors:

mkdir logs
mkdir log_pos

Training & Evaluation

Usage (example):

bash go_adnce.sh $GPU_ID $DATASET_NAME $AUGMENTATION $mode $w1 $w2 $temp

$DATASET_NAME is the dataset name (please refer to https://chrsmrrs.github.io/datasets/docs/datasets/), $GPU_ID is the lanched GPU ID and $AUGMENTATION could be random2, random3, random4 that sampling from {NodeDrop, Subgraph}, {NodeDrop, Subgraph, EdgePert} and {NodeDrop, Subgraph, EdgePert, AttrMask}, seperately. mode represents the model (easy or adnce). w1, w2, temp denote as mu, sigma and temperature respectively.

Commands for reproducing the reported results:

### NCI1
bash go_adnce.sh 0 NCI1 random3 easy 0.7 1.0 0.05
bash go_adnce.sh 0 NCI1 random3 adnce 0.7 1.0 0.05

### PROTEINS
bash go_adnce.sh 0 PROTEINS random4 easy 1.5 1.0 0.05
bash go_adnce.sh 0 PROTEINS random4 adnce 1.5 1.0 0.05

### DD
bash go_adnce.sh 0 DD random2 easy 0.2 1.0 0.2
bash go_adnce.sh 0 DD random2 adnce 0.2 1.0 0.2

### MUTAG
bash go_adnce.sh 0 MUTAG random2 easy 0.7 1.0 0.15
bash go_adnce.sh 0 MUTAG random2 adnce 0.7 1.0 0.15

### COLLAB
bash go_adnce.sh 0 COLLAB random4 easy 0.3 1.0 0.1
bash go_adnce.sh 0 COLLAB random4 adnce 0.3 1.0 0.1

### REDDIT-BINARY
bash go_adnce.sh 0 REDDIT-BINARY random2 easy 0.3 1.0 0.15
bash go_adnce.sh 0 REDDIT-BINARY random2 adnce 0.3 1.0 0.15

### REDDIT-MULTI-5K
bash go_adnce.sh 0 REDDIT-MULTI-5K random2 easy 0.8 1.0 0.15
bash go_adnce.sh 0 REDDIT-MULTI-5K random2 adnce 0.8 1.0 0.15

### IMDB-BINARY
bash go_adnce.sh 0 IMDB-BINARY random2 easy 0.7 1.0 0.55
bash go_adnce.sh 0 IMDB-BINARY random2 adnce 0.7 1.0 0.55

Acknowledgements

The backbone implementation is reference to https://github.com/Shen-Lab/GraphCL/tree/master/unsupervised_TU.