dmlc--dgl
85c9ff01c9
* [Example] Finish adaptive sampling * refactor code and use argparse for command-line usage * Update README.md * add node_per_layer command-line argument
Adaptive sampling for graph representation learning
This is dgl implementation of Adaptive Sampling Towards Fast Graph Representation Learning.
The authors' implementation can be found here.
Performance
Test accuracy on cora dataset achieves 0.84 around 250 epochs when sample size is set to 256 for each layer.
Usage
python adaptive_sampling.py --batch_size 20 --node_per_layer 40