"Ei3SSKrSSKJr SSKJr SSKrSSKJr SSK J r SSK J r S/r \"SSS 9r"S S\ \5rg) N)Iterator)TypeVar)Dataset)SamplerDistributedSampler_T_coT) covariantc\rSrSrSrSS\S\S-S\S-S\S\S \S S4S jjrS \ \ 4S jr S \4S jr S\S S4Sjr Srg)raSampler that restricts data loading to a subset of the dataset. It is especially useful in conjunction with :class:`torch.nn.parallel.DistributedDataParallel`. In such a case, each process can pass a :class:`~torch.utils.data.DistributedSampler` instance as a :class:`~torch.utils.data.DataLoader` sampler, and load a subset of the original dataset that is exclusive to it. .. note:: Dataset is assumed to be of constant size and that any instance of it always returns the same elements in the same order. Args: dataset: Dataset used for sampling. num_replicas (int, optional): Number of processes participating in distributed training. By default, :attr:`world_size` is retrieved from the current distributed group. rank (int, optional): Rank of the current process within :attr:`num_replicas`. By default, :attr:`rank` is retrieved from the current distributed group. shuffle (bool, optional): If ``True`` (default), sampler will shuffle the indices. seed (int, optional): random seed used to shuffle the sampler if :attr:`shuffle=True`. This number should be identical across all processes in the distributed group. Default: ``0``. drop_last (bool, optional): if ``True``, then the sampler will drop the tail of the data to make it evenly divisible across the number of replicas. If ``False``, the sampler will add extra indices to make the data evenly divisible across the replicas. Default: ``False``. .. warning:: In distributed mode, calling the :meth:`set_epoch` method at the beginning of each epoch **before** creating the :class:`DataLoader` iterator is necessary to make shuffling work properly across multiple epochs. Otherwise, the same ordering will be always used. Example:: >>> # xdoctest: +SKIP >>> sampler = DistributedSampler(dataset) if is_distributed else None >>> loader = DataLoader(dataset, shuffle=(sampler is None), ... sampler=sampler) >>> for epoch in range(start_epoch, n_epochs): ... if is_distributed: ... sampler.set_epoch(epoch) ... train(loader) Ndataset num_replicasrankshuffleseed drop_lastreturnc@Uc:[R"5(d [S5e[R"5nUc:[R"5(d [S5e[R"5nX2:dUS:a[ SUSUS- S35eXlX lX0lSUl X`l UR(ao[UR 5UR-S:waI[R"[UR 5UR- UR- 5UlO;[R"[UR 5UR- 5UlURUR-UlX@lXPlg)Nz,Requires distributed package to be availablerz Invalid rank z%, rank should be in the interval [0, ])dist is_available RuntimeErrorget_world_sizeget_rank ValueErrorr r repochrlenmathceil num_samples total_sizerr)selfr r rrrrs Y/home/mande/repo/quber/.venv/lib/python3.13/site-packages/torch/utils/data/distributed.py__init__DistributedSampler.__init__BsR  $$&&"#QRR..0L <$$&&"#QRR==?D  4!8v%J>c$,,/$2C2CCqH $yyT\\"T%6%66$:K:KK D  $yyT\\):T=N=N)NOD **T->->>  cUR(at[R"5nURURUR -5 [R "[UR5US9R5nO'[[[UR555nUR(dZUR[U5- nU[U5::aX"SU- nO:X"[R"U[U5- 5-SU- nOUSURn[U5UR:wa%[!S[U5SURS35eX R"URUR$2n[U5UR&:wa%[!S[U5SUR&S35e[)U5$)N) generatorzNumber of indices (z) does not match total_size ()zNumber of subsampled indices (z) does not match num_samples ()rtorch Generator manual_seedrrrandpermrr tolistlistrangerr!rrAssertionErrorrr r iter)r"gindices padding_sizes r#__iter__DistributedSampler.__iter__ks <<!A MM$))djj0 1nnS%6!DKKMG5T\\!234G~~??S\9Ls7|+=L11dii s7|0K&LL!\ /0G w<4?? * %c'l^3PQUQ`Q`Paabc  ))doo8I8IIJ w<4++ + 0W>\]a]m]m\nnop  G}r&cUR$)N)r )r"s r#__len__DistributedSampler.__len__sr&rcXlg)z Set the epoch for this sampler. When :attr:`shuffle=True`, this ensures all replicas use a different random ordering for each epoch. Otherwise, the next iteration of this sampler will yield the same ordering. Args: epoch (int): Epoch number. N)r)r"rs r# set_epochDistributedSampler.set_epochs  r&) r rrr r rrrr!)NNTrF)__name__ __module__ __qualname____firstlineno____doc__rintboolr$rrr6r9r<__static_attributes__r&r#rrs.f$(''Dj'Dj '  '  '' 'R"(5/"H   s t r&)rcollections.abcrtypingrr*torch.distributed distributedrtorch.utils.data.datasetrtorch.utils.data.samplerr__all__rrrFr&r#rNsD $ ,,  4(LLr&