Skip to content

Latest commit

 

History

History
63 lines (52 loc) · 2.59 KB

readme.md

File metadata and controls

63 lines (52 loc) · 2.59 KB

Edge-aware GNN Link Prediction

MRR Metric

You can utilize following codes to do full batch training on MRR link prediction.

python GNN/GNN_Link_MRR.py \
    --use_PLM_node Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_node.pt \
    --use_PLM_edge Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_edge.pt \
    --path Dataset/goodreads_children/LinkPrediction/ \
    --graph_path Dataset/goodreads_children/processed/children.pkl

MRR Metric using neighbor sampling

You can utilize following codes to handle large-scale graph MRR link prediction.

python GNN/GNN_Link_MRR_loader.py \
    --use_PLM_node Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_node.pt \
    --use_PLM_edge Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_edge.pt \
    --path Dataset/goodreads_children/LinkPrediction/ \
    --graph_path Dataset/goodreads_children/processed/children.pkl

HitsK Metric

You can utilize following codes to do full batch training on HitsK link prediction.

python GNN/GNN_Link_HitsK.py \
    --use_PLM_node Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_node.pt \
    --use_PLM_edge Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_edge.pt \
    --path Dataset/goodreads_children/LinkPrediction/ \
    --graph_path Dataset/goodreads_children/processed/children.pkl

HitsK Metric using neighbor sampling

You can utilize following codes to handle large-scale graph HitsK link prediction.

python GNN/GNN_Link_HitsK_loader.py \
    --use_PLM_node Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_node.pt \
    --use_PLM_edge Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_edge.pt \
    --path Dataset/goodreads_children/LinkPrediction/ \
    --graph_path Dataset/goodreads_children/processed/children.pkl

AUC Metric using neighbor sampling

You can utilize following codes to handle large-scale graph AUC link prediction.

python GNN/GNN_Link_AUC.py \
    --use_PLM_node Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_node.pt \
    --use_PLM_edge Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_edge.pt \
    --graph_path Dataset/goodreads_children/processed/children.pkl

Edge-aware GNN Node Classification

python GNN/GNN_Node.py \
    --use_PLM_node Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_node.pt \
    --use_PLM_edge Dataset/goodreads_children/emb/children_bert_base_uncased_512_cls_edge.pt \
    --graph_path Dataset/goodreads_children/processed/children.pkl