NnUNetTrainerLightMUNet has not been used

[AFuJianPeople]

Hello, I'm very sorry to disturb you. I didn't see nnUNetTrainerLightMUNet being used in the code, it has always been nnUNetTrainer. Could you tell me where nnUNetTrainerLightMUNet is being used? Thank you very much.

[alvinkimbowa]

It is passed as an argument during training as mentioned in the README:

nnUNetv2_train DATASET_ID 2d all -tr nnUNetTrainerLightMUNet

By default nnUNetv2_train uses nnUNetTrainer, and if you want to incorporate a custom model such as LightM-UNet, a quick and dirty way is to write a custom trainer that will train using your custom model. All other stages in the pipeline (data preprocessing and augmentation) remain the same, except the model training. Please take a look at the nnUNet documentation - Extending nnUNet.

[AFuJianPeople]

It is passed as an argument during training as mentioned in the README:

nnUNetv2_train DATASET_ID 2d all -tr nnUNetTrainerLightMUNet

By default nnUNetv2_train uses nnUNetTrainer, and if you want to incorporate a custom model such as LightM-UNet, a quick and dirty way is to write a custom trainer that will train using your custom model. All other stages in the pipeline (data preprocessing and augmentation) remain the same, except the model training. Please take a look at the nnUNet documentation - Extending nnUNet.

Thank you very much for your answer. With your help, I have started modifying my own model.