SpikeGCL (Under review)
This is a PyTorch implementation of SpikeGCL from the paper "A Graph is Worth 1-bit Spikes: When Graph Contrastive Learning Meets Spiking Neural Networks".
Environments
- numpy == 1.23.3
- torch == 1.8+cu111
- torch-cluster == 1.6.1
- torch_geometric == 2.3.0
- torch-scatter == 2.1.1
- torch-sparse == 0.6.17
- CUDA 11.1
- cuDNN 8.0.5
Reproduction
- Cora
python main.py --dataset Cora --threshold 5e-4 --outs 4 --bn --T 64
- Citeseer
python main.py --dataset Citeseer --threshold 5e-3 --bn --T 64
- Pubmed
python main.py --dataset Pubmed --threshold 5e-2 --bn --T 32
- Computers
python main.py --dataset Computers --threshold 5e-2 --outs 32 --bn --T 25
- Photo
python main.py --dataset Photo --threshold 5e-2 --T 15 --bn --outs 8
- CS
python main.py --dataset CS --threshold 5e-1 --outs 32 --T 60 --dropout 0. --bn
- Physics
python main.py --dataset Physics --T 25 --outs 16 --bn --margin 1.0 --threshold 5e-2
- Ogbn-arXiv
python main.py --dataset ogbn-arxiv --T 15 --outs 32 --threshold 5e-2 --bn
- Ogbn-MAG
python main.py --dataset ogbn-mag --T 15 --outs 32 --threshold 5e-3 --bn