Multi-encoder nnU-Net outperforms Transformer models with self-supervised pretraining

ViT
Re-assign community