Expand description
Graph Neural Network kernels.
This module provides GPU-accelerated GNN algorithms:
- GNNInference - Message passing neural network inference
- GraphAttention - Graph attention network (GAT) layers
Structs§
- GATResult
- Result of graph attention.
- GATWeights
- Attention weights for GAT layer.
- GNNConfig
- Configuration for GNN inference.
- GNNInference
- GNN Inference kernel.
- GNNResult
- Result of GNN inference.
- GNNWeights
- GNN layer weights (simulated).
- Graph
Attention - Graph Attention kernel.
- Graph
Attention Config - Configuration for graph attention.
Enums§
- Activation
Type - Activation function type.
- Aggregation
Type - Message aggregation type.