Why methods in __init__ are executed for twice？

[forestbat]

I have a trainer which is like code below：

def __init__(self,  pre_model=None):
        self.device_num = [1, 2]
        self.fab = lightning.Fabric(devices=self.device_num, num_nodes=1, strategy='ddp')
        self.fab.launch()
        with self.fab.rank_zero_first():
            if cfgs["training_cfgs"]["train_mode"]:
                self.traindataset = self.make_dataset("train")
                if cfgs["data_cfgs"]["t_range_valid"] is not None:
                    self.validdataset = self.make_dataset("valid")
            self.testdataset: BaseDataset = self.make_dataset("test")
        print(f"Torch is using {str(self.device)}")

def make_dataset(self, is_tra_val_te: str):
        if (is_tra_val_te == "train") & (hasattr(self, 'traindataset')):
            return self.traindataset
        elif (is_tra_val_te == "valid") & (hasattr(self, 'validdataset')):
            return self.validdataset
        elif (is_tra_val_te == "test") & (hasattr(self, 'testdataset')):
            return self.testdataset
        else:
            data_cfgs = self.cfgs["data_cfgs"]
            # There are many I/Os in GNNDataset.__init__()
            dataset = GNNDataset(data_cfgs, is_tra_val_te)
            return dataset

However make_dataset is still executed for twice (in process 0 and 1), so I want to know, how to solve this problem?