distributed.py

import math
import pickle
import torch
import torch.utils.data as data
import torch.distributed as dist

from torch.utils.data.sampler import Sampler, BatchSampler

__all__ = ['get_world_size', 'get_rank', 'synchronize', 'is_main_process',
           'all_gather', 'make_data_sampler', 'make_batch_data_sampler',
           'reduce_dict', 'reduce_loss_dict']


# reference: https://github.com/facebookresearch/maskrcnn-benchmark/blob/master/maskrcnn_benchmark/utils/comm.py
def get_world_size():
    if not dist.is_available():
        return 1
    if not dist.is_initialized():
        return 1
    return dist.get_world_size()


def get_rank():
    if not dist.is_available():
        return 0
    if not dist.is_initialized():
        return 0
    return dist.get_rank()


def is_main_process():
    return get_rank() == 0


def synchronize():
    """
    Helper function to synchronize (barrier) among all processes when
    using distributed training
    """
    if not dist.is_available():
        return
    if not dist.is_initialized():
        return
    world_size = dist.get_world_size()
    if world_size == 1:
        return
    dist.barrier()


def all_gather(data):
    """
    Run all_gather on arbitrary picklable data (not necessarily tensors)
    Args:
        data: any picklable object
    Returns:
        list[data]: list of data gathered from each rank
    """
    world_size = get_world_size()
    if world_size == 1:
        return [data]

    # serialized to a Tensor
    buffer = pickle.dumps(data)
    storage = torch.ByteStorage.from_buffer(buffer)
    tensor = torch.ByteTensor(storage).to("cuda")

    # obtain Tensor size of each rank
    local_size = torch.IntTensor([tensor.numel()]).to("cuda")
    size_list = [torch.IntTensor([0]).to("cuda") for _ in range(world_size)]
    dist.all_gather(size_list, local_size)
    size_list = [int(size.item()) for size in size_list]
    max_size = max(size_list)

    # receiving Tensor from all ranks
    # we pad the tensor because torch all_gather does not support
    # gathering tensors of different shapes
    tensor_list = []
    for _ in size_list:
        tensor_list.append(torch.ByteTensor(size=(max_size,)).to("cuda"))
    if local_size != max_size:
        padding = torch.ByteTensor(size=(max_size - local_size,)).to("cuda")
        tensor = torch.cat((tensor, padding), dim=0)
    dist.all_gather(tensor_list, tensor)

    data_list = []
    for size, tensor in zip(size_list, tensor_list):
        buffer = tensor.cpu().numpy().tobytes()[:size]
        data_list.append(pickle.loads(buffer))

    return data_list


def reduce_dict(input_dict, average=True):
    """
    Args:
        input_dict (dict): all the values will be reduced
        average (bool): whether to do average or sum
    Reduce the values in the dictionary from all processes so that process with rank
    0 has the averaged results. Returns a dict with the same fields as
    input_dict, after reduction.
    """
    world_size = get_world_size()
    if world_size < 2:
        return input_dict
    with torch.no_grad():
        names = []
        values = []
        # sort the keys so that they are consistent across processes
        for k in sorted(input_dict.keys()):
            names.append(k)
            values.append(input_dict[k])
        values = torch.stack(values, dim=0)
        dist.reduce(values, dst=0)
        if dist.get_rank() == 0 and average:
            # only main process gets accumulated, so only divide by
            # world_size in this case
            values /= world_size
        reduced_dict = {k: v for k, v in zip(names, values)}
    return reduced_dict


def reduce_loss_dict(loss_dict):
    """
    Reduce the loss dictionary from all processes so that process with rank
    0 has the averaged results. Returns a dict with the same fields as
    loss_dict, after reduction.
    """
    world_size = get_world_size()
    if world_size < 2:
        return loss_dict
    with torch.no_grad():
        loss_names = []
        all_losses = []
        for k in sorted(loss_dict.keys()):
            loss_names.append(k)
            all_losses.append(loss_dict[k])
        all_losses = torch.stack(all_losses, dim=0)
        dist.reduce(all_losses, dst=0)
        if dist.get_rank() == 0:
            # only main process gets accumulated, so only divide by
            # world_size in this case
            all_losses /= world_size
        reduced_losses = {k: v for k, v in zip(loss_names, all_losses)}
    return reduced_losses


def make_data_sampler(dataset, shuffle, distributed):
    if distributed:
        return DistributedSampler(dataset, shuffle=shuffle)
    if shuffle:
        sampler = data.sampler.RandomSampler(dataset)
    else:
        sampler = data.sampler.SequentialSampler(dataset)
    return sampler


def make_batch_data_sampler(sampler, images_per_batch, num_iters=None, start_iter=0):
    batch_sampler = data.sampler.BatchSampler(sampler, images_per_batch, drop_last=True)
    if num_iters is not None:
        batch_sampler = IterationBasedBatchSampler(batch_sampler, num_iters, start_iter)
    return batch_sampler


# Code is copy-pasted from https://github.com/facebookresearch/maskrcnn-benchmark/blob/master/maskrcnn_benchmark/data/samplers/distributed.py
class DistributedSampler(Sampler):
    """Sampler that restricts data loading to a subset of the dataset.
    It is especially useful in conjunction with
    :class:`torch.nn.parallel.DistributedDataParallel`. In such case, each
    process can pass a DistributedSampler instance as a DataLoader sampler,
    and load a subset of the original dataset that is exclusive to it.
    .. note::
        Dataset is assumed to be of constant size.
    Arguments:
        dataset: Dataset used for sampling.
        num_replicas (optional): Number of processes participating in
            distributed training.
        rank (optional): Rank of the current process within num_replicas.
    """

    def __init__(self, dataset, num_replicas=None, rank=None, shuffle=True):
        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.shuffle = shuffle

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

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

        # subsample
        offset = self.num_samples * self.rank
        indices = indices[offset: offset + self.num_samples]
        assert len(indices) == self.num_samples

        return iter(indices)

    def __len__(self):
        return self.num_samples

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


class IterationBasedBatchSampler(BatchSampler):
    """
    Wraps a BatchSampler, resampling from it until
    a specified number of iterations have been sampled
    """

    def __init__(self, batch_sampler, num_iterations, start_iter=0):
        self.batch_sampler = batch_sampler
        self.num_iterations = num_iterations
        self.start_iter = start_iter

    def __iter__(self):
        iteration = self.start_iter
        while iteration <= self.num_iterations:
            # if the underlying sampler has a set_epoch method, like
            # DistributedSampler, used for making each process see
            # a different split of the dataset, then set it
            if hasattr(self.batch_sampler.sampler, "set_epoch"):
                self.batch_sampler.sampler.set_epoch(iteration)
            for batch in self.batch_sampler:
                iteration += 1
                if iteration > self.num_iterations:
                    break
                yield batch

    def __len__(self):
        return self.num_iterations


if __name__ == '__main__':
    pass