DistributedEvalSampler.py

import torch.distributed as dist
import torch
from torch.utils.data import Sampler

class DistributedEvalSampler(Sampler):
    def __init__(self, dataset, num_replicas=None, rank=None, shuffle=False, seed=0):
        if num_replicas is None:
            if not dist.is_available():
                raise RuntimeError("Requires distributed package to be available")
            num_replicas = dist.get_world_size()
        if rank is None:
            if not dist.is_available():
                raise RuntimeError("Requires distributed package to be available")
            rank = dist.get_rank()
        self.dataset = dataset
        self.num_replicas = num_replicas
        self.rank = rank
        self.epoch = 0
        # self.num_samples = int(math.ceil(len(self.dataset) * 1.0 / self.num_replicas))
        # self.total_size = self.num_samples * self.num_replicas
        self.total_size = len(self.dataset)         # true value without extra samples
        indices = list(range(self.total_size))
        start_indice = int(self.total_size * self.rank/self.num_replicas)
        end_indice = int(self.total_size * (self.rank+1)/self.num_replicas)
        if start_indice % 5 != 0:
            start_indice = start_indice - start_indice % 5
        if end_indice % 5 != 0:
            end_indice = end_indice - end_indice % 5  
        self.indices = indices[start_indice:end_indice]
        self.num_samples = len(indices)             # true value without extra samples

        self.shuffle = shuffle
        self.seed = seed

    def __iter__(self):
        if self.shuffle:
            # deterministically shuffle based on epoch and seed
            g = torch.Generator()
            g.manual_seed(self.seed + self.epoch)
            indices = torch.randperm(len(self.dataset), generator=g).tolist()
        else:
            indices = list(range(len(self.dataset)))

        # # add extra samples to make it evenly divisible
        # indices += indices[:(self.total_size - len(indices))]
        # assert len(indices) == self.total_size

        # subsample
        indices = self.indices
        # assert len(indices) == self.num_samples

        return iter(indices)

    def __len__(self):
        return self.num_samples

    def set_epoch(self, epoch):
        self.epoch = epoch