In this project:

- 'dcbm&sbm_train' file is used to train the model on the DCBM and SBM generated training sets.
- 'latent_space_model_train' files is used to train the model on the latent space model generated training sets.
- 'sbm_train' files is used to train the model on the SBM dataset generated training sets.
- The folder `real_dataset` contains the real-world data files. You can run `real_data.py` to reproduce the results on the real data.
- Running test_dcbm.py and test_lsm_model.py in the test_3_model folder produces simulation results across different test sets.

In every file:
- logistic_reg.py is responsible for generating the corresponding dataset, building the model, and training the model.
- data_generator_dcsbm.py is used to generate the DCSBM + SBM training and validation dataset.
- data_generator_sbm.py is used to generate the SBM training and validation dataset.
- data_generator_lsm.py is used to generate the LSM training and validation dataset.
- load.py is used to compute the operators needed to train the first-period GNN
- load_local_refinement is used to computer the operators need to train the second-priod GNN


