How could I shuffle train index in every epoch with BatchSampler（DistributedSampler）

[xuanyang19]

How could I call set_epoch() on DistributedSampler in pytorch lightning? Although I set distributedsampler(shuffle = True), the train index is always the same at every epoch. What should I do to make the train index become different at every epoch?
my dataloader:
DataLoader(self.trainset,sampler=BatchSampler(DistributedSampler(self.trainset, shuffle=True), batch_size=self.batch_size,drop_last=False), num_workers=self.num_workers)

[rohitgr7]

it's not required. It will be done automatically for you within Lightning Trainer.