How to use the nnunet.training.network_training.nnUNetTrainer.nnUNetTrainer function in nnunet

from nnunet.training.loss_functions.dice_loss import GDiceLossV2
# from nnunet.training.loss_functions.dice_loss import GDiceLoss
from nnunet.training.network_training.nnUNetTrainer import nnUNetTrainer
from nnunet.utilities.nd_softmax import softmax_helper


class nnUNetTrainer_GDice(nnUNetTrainer):
    def __init__(self, plans_file, fold, output_folder=None, dataset_directory=None, batch_dice=True, stage=None,
                 unpack_data=True, deterministic=True, fp16=False):
        super().__init__(plans_file, fold, output_folder, dataset_directory, batch_dice, stage,
                                              unpack_data, deterministic, fp16)
        self.apply_nonlin = softmax_helper
        self.loss = GDiceLossV2(apply_nonlin=self.apply_nonlin, smooth=1e-5)

from nnunet.training.loss_functions.dice_loss import DC_and_topk_loss
# from nnunet.training.network_training import nnUNetTrainerCE
from nnunet.training.network_training.nnUNetTrainer import nnUNetTrainer


class nnUNetTrainer_DiceTopK10(nnUNetTrainer):
    def __init__(self, plans_file, fold, output_folder=None, dataset_directory=None, batch_dice=True, stage=None,
                 unpack_data=True, deterministic=True, fp16=False):
        super().__init__(plans_file, fold, output_folder, dataset_directory, batch_dice, stage,
                                              unpack_data, deterministic, fp16)
        k = 10
        self.loss = DC_and_topk_loss({'batch_dice':self.batch_dice, 'smooth':1e-5,
        	'do_bg':False}, {'k':k})

from nnunet.training.loss_functions.dice_loss import ExpLog_loss
from nnunet.training.network_training.nnUNetTrainer import nnUNetTrainer
import torch


class nnUNetTrainer_ExpLog(nnUNetTrainer):
    def __init__(self, plans_file, fold, output_folder=None, dataset_directory=None, batch_dice=True, stage=None,
                 unpack_data=True, deterministic=True, fp16=False):
        super().__init__(plans_file, fold, output_folder, dataset_directory, batch_dice, stage,
                                              unpack_data, deterministic, fp16)
        self.weight = torch.cuda.FloatTensor([0.2,0.8])
        self.loss = ExpLog_loss({'batch_dice': self.batch_dice, 'smooth': 1e-5, 'do_bg': False, 'square': False}, {'weight':self.weight})