MIC-DKFZ / nnUNet

Apache License 2.0
5.95k stars 1.77k forks source link

Pretrained and continous training #2609

Open gogochen07 opened 2 days ago

gogochen07 commented 2 days ago

Hi, I would like to leverage the model from dataset 001 to further train on it using dataset 002, I tried use the -pretrained_weights flag, but it shows the network architecture is not consistent. Could I know how to achieve the further finetuning? Thanks!