An unofficial implementation of Graph Transformer:
Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification) - IJCAI 2021 > https://www.ijcai.org/proceedings/2021/0214.pdf
This GNN architecture is implemented based on Section 3.1 (Graph Transformer) in the paper.
I implemented the code by referring to this repository, but with some modifications to match with the original published paper in IJCAI 2021.
pip install graph-transformerimport torch
from graph_transformer import GraphTransformerModel
model = GraphTransformerModel(
node_dim = 512,
edge_dim = 512,
num_blocks = 3, # number of graph transformer blocks
num_heads = 8,
last_average=True, # wether to average or concatenation at the last block
model_dim=None # if None, node_dim will be used as the dimension of the graph transformer block
)
nodes = torch.randn(1, 128, 512)
edges = torch.randn(1, 128, 128, 512)
adjacency = torch.ones(1, 128, 128)
nodes = model(nodes, edges, adjacency)Note: If your graph does not have edge features, you can set edge_dim and edges (in the forward pass) to None.