Getting familiar with Graph Neural Nets and the pytorch_geometric library
- Set up a python3 conda environment, then
pip install -r requirements.txt
- Conda install pytorch as specified here
- Install pytorch_geometric with
./install_pytorch_geometric.sh
, or one-by-one as specified here - Run 'main.py' out of the directory
./source
For a given random geometric graph in a plane, predict the number of neighbors within a threshold distance θ. In particular:
- Randomly sample N nodes in a unit square.
- Connect nodes within a threshold distance θ_max.
- Label each node with the number of direct neighbors within distance θ < θ_max.
- Train a GNN to predict the label of each node.
- To install MongoDB on MacOS, execute
brew install mongodb
, then start it as a service withbrew services start mongodb
- Start mongodb with
mongodb
, you'll be in the MongoDB shell - In there, set up a new database called sacred with
use sacred
- Maybe: Insert a dummy entry with
db.movie.insert({"name":"tutorials point"})
- To set up Omniboard, follow the steps here
- Execute
omniboard -m <host>:27017:sacred
to start up Omniboard atlocalhost:9500