This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from scipy.stats import entropy | |
import numpy as np | |
def label_informativeness(edge_index, labels): | |
''' | |
Computes the label informativeness metric proposed in | |
Platonov et al. (2022) | |
Characterizing Graph Datasets for Node Classification: | |
Beyond Homophily-Heterophily Dichotomy | |
https://arxiv.org/abs/2209.06177 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
''' | |
Wandb Sweep: Multiple seeds same config/parameters | |
Script to run wandb sweeps with a number of different seeds: | |
Given some seeds, this sweep controller samples a model configuration and starts one run per seed with the exact same setup. | |
I made this for a model that had a high variance to get somewhat reliable values for every configuration. | |
In order to group the runs with the different seeds in the wandb webapp, an additional value 'sweep_step_id' is passed | |
that is the same for each configuration. | |
To execute, run |