Python=3.10.14
torch==2.2.2
torch_geometric==2.5.2
numpy==1.24.3
pytz==2024.1
sklearn==1.4.2
- Run
train_N_M.py
. You can adjust parameters intrain_N_M.py
. It's currently set to train forN=20, M=3
andN=50, M=5
problems. The hyperparameters are configured as described in the paper, maintaining consistency with the original research.
- Run
test_N.py
. Choose the model to test by modifying the parameter intest_N.py
(seed: 1235). It's set to use the trained model (N=20, M=3
andN=50, M=5
) in the result folder, but you can switch to a model you've trained.
- NeuroGuide's code execution is based on the POMO. We thank them for their contribution.