Originaly coding credit to Huang Xin, Xiaoyun continued contributing it.
1 Donwload dataset
wget https://s3.us-east-2.amazonaws.com/rapidsai-data/cugraph/test/datasets.tgz
2 extract the dataset
tar -xvzf datasets.tgz
3 rename the dataset
mv datasets data
4 To run the cugraph random walk, please use
python RW_cugraph_benchmark.py
5 you will get a bunch of .csv files.
6 to run DGL sampling code, you need to install pytorch and DGL inside of the RAPIDS container.
The container is here.
https://hub.docker.com/r/rapidsai/rapidsai/
Then go into the container and run the pip conmands to install pytorch and DGL.
pip3 install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html
pip install dgl-cu111 -f https://data.dgl.ai/wheels/repo.html
7 run DGL random walk, it is similar to step 4.
python RW_DGL_benchmark.py