I want to train your architecture using my own dataset. I already preprocessed the dataset using nnunet and made a config.yaml file using the examples you provided for the three datasets here. I set the hdfs_base parameter to Task511_3DTransUNet_encoder_only with Task511 being my own dataset. However, from the code in train.py I see that this hdfs_base parameter is only used in get_default_configuration() for the name of the folder. Is that the only parameter I need to set in order to train the three different configurations - encoder only, decoder only and encoder plus decoder?
Hello,
I want to train your architecture using my own dataset. I already preprocessed the dataset using nnunet and made a config.yaml file using the examples you provided for the three datasets here. I set the
hdfs_baseparameter toTask511_3DTransUNet_encoder_onlywithTask511being my own dataset. However, from the code intrain.pyI see that thishdfs_baseparameter is only used inget_default_configuration()for the name of the folder. Is that the only parameter I need to set in order to train the three different configurations - encoder only, decoder only and encoder plus decoder?