The configurations in legacy/ are designed for torchdistill <= v0.3.3 with torchvision =< v0.13.1.

The results obtained with the configurations in official/ are confirmed and reported in papers, thus you could use them and refer to the results as baselines in your study.

Most of the configurations in sample/ should be used for debugging or as starter configs to design your experiments. If you use the methods as baselines, you SHOULD tune the hyperparameters in the configuration for a fair comparison.