How to use sampler on distributed system by pytorch

I define a kind of sampler by myself, and I want to training the model on distributed system. I do not know how to modify the code to adopt to distributed system. The following is the code of sampler.

import torch
import numpy as np

class CategoriesSampler():

    def __init__(self, label, n_batch, n_cls, n_per):
        self.n_batch = n_batch
        self.n_cls = n_cls
        self.n_per = n_per

        label = np.array(label)
        self.m_ind = []
        for i in range(max(label) + 1):
            ind = np.argwhere(label == i).reshape(-1)
            ind = torch.from_numpy(ind)


    def __len__(self):
        return self.n_batch
    def __iter__(self):
        for i_batch in range(self.n_batch):
            batch = []
            # print(len(self.m_ind))
            classes = torch.randperm(len(self.m_ind))[:self.n_cls]
            for c in classes:

                l = self.m_ind[c]

                pos = torch.randperm(len(l))[:self.n_per]
                # else:
                #     print(c)

            batch = torch.stack(batch).t().reshape(-1)
            yield batch

You could adapt or derive from DistributedRandomSampler.
There is nothing really special going on besides the splitting of the indices.