nnUNets raw dataset and trained models stored in the conventional format: https://github.com/MIC-DKFZ/nnUNet/blob/master/documentation/how_to_use_nnunet.md

Raw datasets used for training are in the folder raw/<dataset-name>
Splits used during training for each dataset are in preprocessed/<dataset-name>/splits_final.json
Best checkpoints are in results/<dataset-name>/<training-plan>/<fold_#>/checkpoint_best.pth